Physiological Consequences of Environmental Contamination in an Elasmobranch with Matrotrophic Histotrophy, the Round Stingray (Urobatis Halleri)

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2018-01-15 Physiological consequences of environmental contamination in an elasmobranch with matrotrophic histotrophy, the Round Stingray (Urobatis halleri)

Lyons, Katherine

Lyons, K. (2018). Physiological consequences of environmental contamination in an elasmobranch with matrotrophic histotrophy, the Round Stingray (Urobatis halleri) (Unpublished doctoral thesis). University of Calgary, Calgary, AB. http://hdl.handle.net/1880/106331 doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Physiological consequences of environmental contamination in an elasmobranch with

matrotrophic histotrophy, the Round Stingray (Urobatis halleri)

Katherine Danielle Hohman Lyons

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN BIOLOGICAL SCIENCES

CALGARY, ALBERTA

JANUARY, 2018

A range of physiological biomarkers were compared in two populations of Round Stingray

(Urobatis halleri) that differed primarily in PCB exposure. Females, and their embryos, were sampled each month of pregnancy at both sites, while adult males were matched with a 40-day subset of females to investigate the effect of sex and its interaction with PCB exposure. I hypothesized that exposure would have negative energetic outcomes for PCB-exposed, compared to reference population, stingrays. Adverse impacts were widespread. Number of offspring was not affected by PCB exposure; however, exposure exacerbated maternal tissue mass, and quality, loss during pregnancy. PCB-exposed females also compromised osmoregulation with urea, an important osmolyte, reduced in maternal plasma. Higher liver quality in contaminant-exposed females was not associated with higher embryo quality, suggesting contaminants increase metabolic demands in adult females and lead to inefficiencies in embryos’ use of maternal resources. Embryos also influenced their uterine environment, as they were steroidogenic and capable of osmoregulating very early in development. I found sex- related differences in embryo mass only at the reference site, suggesting that contaminant effects on males begin in utero. These contaminant sex-related effects extended into adulthood, as relative liver mass and energy content were lower in comparably-sized adult males than females, whereas fewer differences were found between adults at the reference site. Higher energy generation potential, combined with lower tissue quality in contaminant-exposed adult males suggests they are more energetically compromised than females, despite the latter’s costly pregnancy demands. Regardless of sex, contaminant exposure had negative impacts on the ability of adult stingrays to mount a robust secondary stress response as reflected by lower plasma glucose levels after stress, thus potentially impairing their ability to respond to acute

ii stressors. Effects found both in utero and in adulthood suggest that contaminants have a significant, and potentially life-long, impact on Round Stingray homeostasis. This has implications for other species with greater contaminant burdens. Contaminant exposure, and its interactive effects with sex and age, should be included as part of effective elasmobranch management.

iii

Acknowledgements

First and foremost, I need to thank my family and friends for all of the unwavering support they provided to me while obtaining this degree. I could not have overcome all the struggles I went through without these people in my life. Specifically, and in no particular order,

I want to thank my friends Ryan Freedman, Chris Bedore, Ramya Singh, Noreen Singh, Joanne

Hou, Carrie Espasandin, “Arthur Cheng”, Brendan Cooper, Joe Bizzarro and my family, mom, dad and sister Rachael, who were there with me through the worst of it. I need to give a shout out to Ralph Appy who was out in the field with me every single fishing day with a smile and

Dennis Noesen who generously shuttled us between the mainland and Catalina Island. Thanks to

Chris Lowe for providing me lab space and storage of my samples while in the field.

I need to give a mountain of gratitude to my advisor, Dr. Wynne-Edwards, for throwing me a life ring when I thought this was a sinking ship. In the much too short time Dr. Wynne-

Edwards advised me, I learned so much more in four months than I did in the previous four years. Dr. Wynne-Edwards restored my faith in my science and played an incredibly instrumental role in helping me write and analyze my thesis into the strongest product it could be. I will be forever appreciative and grateful for the advice, guidance and mentorship that Dr.

Wynne-Edwards has given to me.

iv Dedication

To the Round Stingray, whose sacrifice furthered our scientific understanding

v Table of Contents

Abstract ...... ii Acknowledgements ...... iv Dedication ...... v Table of Contents ...... vi List of Tables ...... x List of Figures and Illustrations ...... xi List of Symbols, Abbreviations and Nomenclature ...... xv

Chapter 1: Introduction ...... 1 1.1 Elasmobranch Physiology ...... 1 1.2 Polychlorinated Biphenyl Contaminants ...... 3 1.3 Patterns of Accumulation ...... 4 1.4 Organochlorine Effects in Elasmobranchs ...... 5 1.5 Knowledge Gaps ...... 7 1.6 Round Stingray Ecology in Southern California ...... 7 1.7 Sources of Contamination at the Mainland Site ...... 8 1.8 Comparisons with the Offshore Site ...... 10 1.9 Research Objectives ...... 12

Chapter 2: Physiological Consequences of Legacy PCB Contamination in an Elasmobranch with Matrotrophic Histotrophy, the Round Stingray (Urobatis halleri): Impaired Intrauterine Development ...... 13 2.1 Introduction ...... 13 2.2 Methods ...... 15 2.2.1 Study sites...... 15 2.2.2 Sampling...... 17 2.2.3 Data analysis...... 17 2.2.3.1 Environmental factors...... 17 2.2.3.2 PCB quantification...... 17 2.2.3.3 Fecundity...... 19 2.2.3.4 Developmental stage...... 19 2.2.3.5 Sex-specific effects...... 19 2.2.3.6 Embryo quality...... 20 2.2.3.7 Intrauterine growth variability...... 20 2.3 Results ...... 20 2.3.1 Timing of ovulation...... 20 2.3.2 Fecundity...... 21 2.3.3 Developmental stage...... 21 2.3.4 Sex-specific effects...... 24 2.3.5 Embryo quality...... 24 2.3.6 Intrauterine growth variability...... 27 2.4 Discussion ...... 30 2.4.1 Fecundity...... 30 2.4.2 Developmental stage...... 31 2.4.3 Sex-specific effects ...... 31

vi 2.4.4 Embryo quality...... 34 2.4.5 Intrauterine competition...... 34 2.4.6 Conclusions...... 35

Chapter 3: Developmental Embryonic Steroid Contributions to Histotroph in Round Stingrays ...... 37 3.1 Introduction ...... 37 3.2 Methods ...... 39 3.2.1 Study populations...... 39 3.2.2 Sampling...... 39 3.2.3 Hormone quantitation...... 40 3.2.3.1 Chemicals and reagents...... 40 3.2.3.2 Sample preparation procedure...... 41 3.2.3.3 LC-ESI/MRM analysis...... 41 3.2.3.4 Quantitation method...... 41 3.2.4 Data analysis...... 42 3.3 Results ...... 43 3.3.1 Males...... 44 3.3.2 Maternal Plasma...... 44 3.3.3 Histotroph...... 47 3.3.4 Maternal vs. embryo...... 53 3.4 Discussion ...... 56 3.4.1 Adult plasma sex steroids...... 56 3.4.2 Mothers: hormone sink or source? ...... 57 3.4.3 Yolk as a source...... 57 3.4.4 Embryos as a source...... 58 3.4.5 Role of uterine hormones...... 59 3.4.6 Litter sex composition...... 60 3.4.7 Conclusions...... 60

Chapter 4: Physiological Consequences of Environmental PCB Contamination in an Elasmobranch with Matrotrophic Histotrophy, the Round Stingray (Urobatis halleri): Attenuation of the Acute Stress Response ...... 62 4.1 Introduction ...... 62 4.2 Methods ...... 64 4.2.1 Sites...... 64 4.2.2 Sampling...... 66 4.2.3 Assays...... 67 4.2.3.1 Primary response: relative 1α-OH-corticosterone ...... 67 4.2.3.2 Secondary response: energy substrates...... 68 4.2.4 Data analysis...... 69 4.3 Results ...... 69 4.3.1 Females...... 69 4.3.1.1 Baseline measures...... 69 4.3.1.2 Primary stress response...... 70 4.3.1.3 Secondary stress response...... 71 4.3.1.4 Tissue quality...... 73

vii 4.3.2 Males versus Females...... 73 4.3.2.1 Sex effects within the contaminant-exposed site...... 73 4.3.2.2 Site effects within the stress response between sites...... 74 4.4 Discussion ...... 78 4.4.1 Conclusions ...... 82

Chapter 5: Physiological Consequences of PCB Contamination in an Elasmobranch with Matrotrophic Histotrophy, the Round Stingray (Urobatis halleri): Reduced Concentrations of the Primary Osmolyte, Urea ...... 83 5.1 Introduction ...... 83 5.2 Methods ...... 85 5.2.1 Field Sites ...... 85 5.2.2 Sampling ...... 86 5.2.3 Assays...... 87 5.2.4 Data analysis...... 89 5.3 Results ...... 89 5.3.1 Tissue condition...... 90 5.3.1.1 Mothers...... 90 5.3.1.2 Embryos...... 90 5.3.2 Fluid osmolytes...... 92 5.3.2.1 Maternal plasma...... 92 5.3.2.2 Histotroph...... 92 5.3.3 Muscle Osmolytes...... 95 5.3.3.1 Mothers...... 95 5.3.3.2 Embryos...... 95 5.3.4 Urea metabolism...... 97 5.3.4.1 Mothers...... 97 5.3.4.2 Embryos...... 97 5.4 Discussion ...... 104 5.4.1 Plasma vs. histotroph...... 104 5.4.2 Embryo osmoregulatory capability...... 105 5.4.3 Contaminant effects...... 107 5.4.4 Conclusions...... 108

Chapter 6: Physiological Consequences of PCB Contamination in an Elasmobranch with Matrotrophic Histotrophy, the Round Stingray (Urobatis halleri): Exposure Differentially Affects Males More than Females by Lowering Tissue Quality and Increasing Metabolic Capacity ...... 110 6.1 Introduction ...... 110 6.2 Methods ...... 112 6.2.1 Study Sites ...... 112 6.2.2 Sampling ...... 112 6.2.3 Tissue quality...... 113 6.2.4 Metabolic capacity...... 113 6.2.5 Data analysis...... 114 6.3 Results ...... 115 6.3.1 Effect of pregnancy...... 115

viii 6.3.1.1 Tissue quality...... 118 6.3.1.2 Metabolic capacity...... 119 6.3.2 Adult sex differences...... 120 6.3.2.1 Tissue quality...... 120 6.3.2.2 Metabolic capacity...... 122 6.3.3 Embryo metabolism...... 124 6.3.3.1 Tissue quality...... 124 6.3.3.2 Metabolic capacity...... 128 6.3.3.3 Maternal-embryo tradeoffs...... 128 6.4 Discussion ...... 129 6.4.1 Effect of pregnancy...... 129 6.4.2 Adult sex differences...... 130 6.4.3 Embryo metabolism...... 133 6.4.4 Conclusions...... 135

Chapter 7: Perspectives and Future Directions ...... 136 7.1 Project Summary ...... 136 7.1.1 Reproductive impairment in females...... 136 7.1.1.1 Pregnancy was expensive...... 136 7.1.1.2 Contaminant-exposure was energetically compromising...... 136 7.1.1.2 The acute stress response was impaired...... 137 7.1.2 Adult males have impairments from contaminant exposure as well...... 137 7.1.2.1 Sex influences physiology...... 137 7.1.2.2 Contaminant exposure impaired male homeostasis...... 138 7.1.3 Embryos influenced their uterine environment and competed for maternal resources...... 138 7.1.3.1 Embryonic steroids and osmoregulatory capacity were highly developed...... 138 7.1.3.2 Embryos competed for resources...... 139 7.1.4 Maternal contaminant exposure affects her developing embryos...... 139 7.1.4.1 Contaminants impaired embryonic resource utilization...... 139 7.1.4.2 Embryo sex matters...... 139 7.1.5 Conclusions...... 140 7.2 Future Directions ...... 140

References ...... 145

ix List of Tables

Table 2.1. Organochlorine concentrations in a subset of stingrays ...... 18

Table 3.1. Mass transitions and limits of quantitation for all steroids...... 43

Table 3.2. Histotroph Steroids by stage of pregnancy...... 49

Table 4.1. Baseline male stingray plasma stress variable values...... 74

Table 4.2. Summary of sex differences in the response to capture stress...... 77

Table 5.1. Enzyme activities across pregnancy in mothers and embryos...... 99

Table 6.1. Tissue changes across pregnancy by site...... 116

Table 6.2. Adult hepatic enzyme activities by site and sex...... 119

Table 6.3. Main effects and interactions for site, sex, and development...... 134

x List of Figures and Illustrations

Figure 2.1. Map of sampling sites ...... 16

Figure 2.2. Female reproductive success ...... 22

Figure 2.3. Embryo growth can be aligned by a developmental marker, clasper day ...... 23

Figure 2.4. Sex and site differences in mass of the largest embryo...... 25

Figure 2.5. Hepatic lipids by sex and site...... 26

Figure 2.6. Mass variability within litters...... 28

Figure 2.7. Within-litter competition for resources...... 29

Figure 2.8. Water temperatures by site...... 33

Figure 3.1. Testosterone concentrations in maternal plasma by pregnancy stage and site...... 46

Figure 3.2. Histotroph progesterone, 17α-OH-progesterone, testosterone and estrogen changes with pregnancy stage...... 50

Figure 3.3. Histotroph steroid changes at the time of external sexual differentiation...... 51

Figure 3.4. Impact of yolk on histotroph progesterone and testosterone concentrations...... 52

Figure 3.5. Associations between maternal plasma and histotroph testosterone...... 54

Figure 3.6. Impact of sex-biased litters on histotroph steroids...... 55

Figure 4.1. Adult male and female sampling locations...... 65

Figure 4.2. Maternal plasma responses to capture stress by site...... 72

Figure 4.3. Sex and site effects on plasma and liver energy reserves...... 76

Figure 5.1. Maternal and embryo muscle condition throughout pregnancy...... 91

Figure 5.2. Maternal histotroph osmolytes...... 94

Figure 5.3. Embryo muscle osmolytes...... 96

Figure 5.4. Embryo enzyme activities over ontogeny...... 98

Figure 5.5. Maternal and embryonic enzyme activity comparisons...... 100

Figure 5.6. Accelerated protein catabolism in contaminant-exposed embryos ...... 103

xi Figure 6.1. Conversion of energy reserves into litter mass...... 117

Figure 6.2. Adult tissue quality measures by site and sex...... 121

Figure 6.3. Adult metabolic capacity by substrate group...... 123

Figure 6.4. Embryo quality by site...... 125

Figure 6.5. Relative hepatic energy reserves...... 127

Figure 6.6. Male mass relative to disk width by site...... 132

Figure 7.1. Elasmobranch PCB burdens for a range of species...... 142

xii

List of Symbols, Abbreviations and Nomenclature

Symbol Definition # number % percent ° degree ~ approximately 11-DHC 11-dehydrocorticosterone AC# Animal care protocol number AcAc acetoacetic acid ACN acetonitrile AhR aryl hydrocarbon receptor ANCOVA analysis of covariance ANOVA analysis of variance ATPase adenosine triphosphatase Aug August B/Bo bound / unbound BOH 3-hydroxybutyric acid C Celcius CA California cm centimeter CV coefficient of variation CYP1A cytochrome P4501A DDT dichlorodiphenyltrichloroethane df degrees of freedom DW disk width DNA deoxyribonucleic acid E exposed site EDTA ethylenediaminetetraacetic acid ELISA enzyme-linked immunosorbent assay EPA Environmental Protection Agency ERDDAP http://coastwatch.pfeg.noaa.gov/erddap/index.html ESI electrospray ionization FFA free fatty acids g gram g standard acceleration due to gravity GLUT2 glucose transporter 2 GSase glutamine synthetase H2O water IS internal standards Jun June kJ kilojoule kJ/dw kilojoule per disk width kJ/g kilojoule per gram

xv km kilometer KOH potassium hydroxide KW Kruskal-Wallis test LC liquid chromatography LC-ESI/MRM liquid chromatography-electrospray ionization/multiple reaction monitoring mass spectrometry LCMS liquid chromatography mass spectrometry LLOQ lower limit of quantification LN natural log LOQ limit of quantification lw lipid weight M-F male-female M-M male-male MANOVA multivariate analysis of variance mg milligram mins minutes mL milliliter mm millimeter mM millimolar MODIS moderate resolution imaging spectroradiometer mOsm/kg milliosmole per kilogram MS methanesulfonate N north Na+/K+ sodium/potassium Na2HPO4 sodium phosphate dibasic NADH nicotinaminde adenine dinucleotide NaF sodium fluoride ng/g nanogram per gram ng/mL nanograms per milliliter NH3 ammonia nmol/L nanomoles per liter nmol/mg nanomols per milligram NOAA National Oceanic and Atmospheric Administration OHP 17α-OH-progesterone OUC ornithine-urea cycle PCB(s) polychlorinated biphenyl(s) PEPCK phosphoenolpyruvate carboxykinase pg/mL pictograms per milliliter pH potential of hydrogen QC quality control R reference site (not to be confused with the software R) rpm revolutions per minute s seconds SBNWR Seal Beach National Wildlife Refuge SD standard deviation

xvi Sept September TCA trichloroacetic acid TCDD 2,3,7,8-tetrachlorodibenzodioxin TMAO trimethylamine-N-oxide urea:TMAO urea to trimethylamine-N-oxide ratio W west ww wet weight

ZnSO4•7H2O zinc sulfate with seven water molecules µg microgram µL microliter µm micrometer µmol ADP produced/mg micromoles adenosine diphosphate produced per milligram protein/h protein per hour US United States µmol/min/g micromoles per minute per gram µmoles micromoles

xvii 1

Chapter 1: Introduction

The focus of this thesis is the comprehensive physiological assessment of adverse effects from legacy environmental exposure of polychlorinated biphenyls (PCBs) in Round Stingrays Urobatis halleri. To place these findings in context, this chapter will review the unique aspects of elasmobranch (sharks, skates and rays) physiology, the known effects of environmental PCB contaminant exposure, patterns of PCB accumulation in marine environments, known effects on elasmobranchs, and knowledge gaps surrounding impacts of PCB contamination. The energetic demands of pregnancy with matrotrophic histotrophy are identified as a critical stress point when effects of environmental PCB contamination should be readily detectable, and two pure seawater sites that are geographically proximate, genetically isolated, and differ primarily on their established differences in PCB contamination of Round Stingrays are compared and contrasted. The chapter concludes with a description of the research strategy, aims, and specific hypotheses to be tested.

1.1 Elasmobranch Physiology Elasmobranchs (sharks, skates, and rays) are a diverse, ancient group of cartilaginous fishes that possess many characteristics that separate them from other vertebrates. Some of their defining features from other fishes include their cartilaginous skeleton, lack of operculum and swim bladder, and modified scales (i.e. dermal denticles). Besides physical features, elasmobranch physiology also differs from teleosts in some important ways, including metabolic fuels, osmoregulation and reproduction. These departures lead to different energetic strategies, and could lead to differences in how animals cope with stressors, with a particular focus in this thesis on environmental anthropogenic chemicals such as polychlorinated biphenyls (PCBs). Elasmobranchs differ from teleost and mammals in the primary fuels extrahepatic tissues, particularly muscles, use for energy. Ketone bodies produced from fatty acid oxidation are the primary energy substrate found in the circulation of elasmobranchs (Moon and Mommsen, 1987; Singer and Ballantyne, 1989). Elasmobranch muscle lacks carnitine palmitoyltransferase activity (enzyme involved in fatty acid metabolism) whereas β-hydroxybutyrate dehydrogenase activity (enzyme involved in ketone body metabolism) is high (Anderson, 1990). Unlike mammals and teleosts, elasmobranchs lack albumin proteins in their blood, which assists in the blood transport of fatty acids. The absence of albumin is thought to be a contributing factor to why fatty acid

2 oxidation is restricted to the liver and use of hydrophilic ketone bodies and amino acids as the major energy source is important for elasmobranchs (Anderson, 1990; Watson and Dickson, 2001). Elasmobranchs also differ in their strategy for dealing with osmoregulation challenges of living in a marine environment. Most marine teleosts produce ammonia as their nitrogenous waste product and remove this toxic waste from the body in a timely manner (Forster and Goldstein, 1969). However, marine fishes are then challenged because they are hypo-osmotic to their environment and are continually losing water to that environment. Elasmobranchs instead convert their nitrogenous waste to urea and maintain relatively high levels of it in circulation to make them slightly hyperosmotic to the surrounding marine environment (Goldstein and Forster, 1971). Since urea is also toxic, elasmobranchs have mitigated this risk by also retaining trimethylamine- N-oxide (TMAO) to prevent protein denaturation in the presence of urea (Goldstein et al., 1967; Yancey and Somero, 1979). Elasmobranchs maintain their high urea through activity of the ornithine urea cycle (Goldstein and Forster, 1971). Interestingly, the activities of these enzymes vary for elasmobranchs that live in marine, freshwater, and those that can move between the haline, euryhaline, and fresh water environments. Other adaptations include modifications of the gills (Fines et al., 2001) and kidneys (Morgan et al., 2003; Schmidt-Nielsen et al., 1972) to reduce urea loss, and development of the rectal gland to assist with the clearance of sodium chloride (Evans et al., 2004; Goldstein and Forster, 1971; Pillans et al., 2005). Reproduction also differs between elasmobranchs and most teleosts. Fertilization takes place internally in elasmobranchs with approximately 43% of species being oviparous and the remainder viviparous (Compagno, 1990; Wourms, 1977; Wourms and Demski, 1993). Among the viviparous species, there is a wide range of matrotrophic strategies employed with co-varying levels of maternal investment. Females that internally gestate but provide no supplemental nutrition are referred to as lecithotrophic. The term matrotrophic is applied to internal gestation with ongoing maternal nutrition (Wourms et al., 1988). Among the matrotrophic elasmobranchs, females employ diverse strategies to provide supplemental energy to their young. Examples range from nutritionally enriched uterine secretions (histotrophy) to placental connections between maternal and embryonic vasculatures, to intrauterine cannibalism (adelphophagy; Wourms et al., 1988). Regardless of the reproductive mode or level of matrotrophic investment, elasmobranch females invest significant resources into young, including even the egg-laying species. At

3 hatching, neonate elasmobranchs are fully formed and completely self-sufficient. Prolonged maternal care does not extend beyond the birth, leading to large offspring investments by viviparous females (Hussey et al., 2010), such that newborns tend to appear similar to the adults, although smaller.

1.2 Polychlorinated Biphenyl Contaminants Organochlorine contaminants are a class of compounds that are man-made for uses ranging from agriculture to industry. Many of these compounds are chemically designed to be resistant to degradation, making them persistent once they are released into the environment. Polychlorinated biphenyls (PCBs) are a class of organic contaminants characterized by two benzyl rings with one to ten chlorine atom constituents. Manufactured as lubricants and industrial coolants, PCBs were commercially produced from 1930 until around 1970 as diverse chemical mixtures (e.g. Arochlor, Clophen). Upon recognition of their ability to cause physiological disruptions, including immune suppression (Schwacke et al., 2011; Serdar et al., 2014), reproductive impairment (Folland et al., 2016; Leijs et al., 2014; Reijnders, 1986) and associations with cancer (Lauby-Secretan et al., 2016; Silberhorn et al., 1990), they were banned in many countries and world-wide production of PCBs was phased out around 1989. However, the robust chemical structure of PCBs has allowed them to persist in the environment decades after cessation of their production. In addition, PCBs are highly lipophilic and tend to accumulate in tissue lipid depots, leading to bio-magnification through food webs. Therefore, animals feeding at high trophic levels tend to accumulate the greatest concentrations of PCB contaminants (Borgå et al., 2012; Gobas and Arnot, 2010; Walters et al., 2011). PCBs have a wide range of effects, causing physiological disruptions at many levels in fishes. Hormone synthesis and, as a consequence, signaling is altered for a variety of hormones including thyroid (Brar et al., 2010; Brown et al., 2004), estrogens (Bugel et al., 2011; Thomas, 1989), androgens (Crain et al., 1998; Feist et al., 2005), and corticosteroids (Hontela and Vijayan, 2008; Quabius et al., 1997), among others. One well-studied disruption is the effect PCBs have on the stress response. The stress response is the cascade of physiological events that results from the perception of a stressor (e.g. predation event, homeostatic imbalance), typified by an increase in plasma glucocorticoids and glucose. An attenuated or reduced stress response can negatively impact an animal’s fitness because the animal can no longer appropriately respond to and cope

4 with the stressor (Vijayan et al., 2010), including removing themselves from the stressor (Marentette et al., 2013). Previous studies have documented diminished stress responses in fishes exposed to PCBs directly or via agonists (Aluru et al., 2004; Wiseman and Vijayan, 2011). Closely tied with hormone disruptions by PCB exposure are alterations to metabolism. PCBs can alter metabolism directly by interfering with thyroid hormone signaling (Duntas, 2015; Tabuchi et al., 2006) or by imposing an additional energetic cost on animals, reducing tissue mass and energy content (Feist et al., 2005; Marentette et al., 2013). These effects have negative implications for fitness. For example, growth was slower in Brook Trout fry (Salaelinus fontinalis; Mauck et al., 1978) and Winter Flounder embryos (Pseudopleuronectes americanus) spawned from adults sampled from contaminated areas (Black et al., 1988). Reductions in fitness can also have cascading effects on reproduction, by reducing fecundity (Bugel et al., 2011; Bursian et al., 2013) or embryo fitness and survival (Roy et al., 2011). Taken together, PCBs have the demonstrated potential to negatively affect individual physiology, with implications for population health.

1.3 Patterns of Accumulation Due to their lipophilic chemical structures, PCB contaminants bioaccumulate in lipid rich tissues and biomagnify towards species at higher trophic levels (Safe, 1994). Bioaccumulation is the accretion of contaminants overtime, while biomagnification is the increase in concentration of contaminants as they are transferred up through the food web. In addition, PCBs, as with other organic contaminants, show differential accumulation patterns between males and females (Binnington and Wania, 2014; Madenjian et al., 2016). Since females directly convert their lipid reserves into offspring through the production of eggs or secondary sources of nutrition (i.e. milk), which are lipid-rich, they passively offload lipophilic contaminants onto their offspring, thereby lowering their own contaminant load (Debier et al., 2003; Lundin et al., 2016). Males do not have similar opportunities to offload contaminants and therefore continually bioaccumulate contaminants throughout their lifetime or until they reach equilibrium (Borrell et al., 1995). As examples, male Orcas (Orcinus orca) continually increase their contaminant concentrations (Ross et al., 2000), whereas Long-Finned Pilot Whales (Globicephala melas) showed no relationship with age and the authors concluded that these males were saturated with contaminants and had reached an equilibrium in which contaminant intake and excretion or metabolism were roughly

5 equal (Borrell et al., 1995). In marine mammals that have secondary care of offspring, females can offload 30 to 80% of their total contaminant load onto their first born (Lee et al., 1996) compared to spawning fishes that offload only 5-25% where maternal nutrition is derived exclusively from the egg yolk (Niimi, 1983). Although female offloading is beneficial to those adult females, it carries costs for their offspring. Female offloading results in early developmental exposure of embryos to xenobiotics. Animals are most vulnerable to environmental influences (both anthropogenic and natural) during development and in their juvenile stages (Birnbaum and Tuomisto, 2000). Therefore, any physiological response to contamination would likely be most noticeable or pronounced at the early life stages. Previous studies have documented effects in a variety of species exposed to PCB contaminants during development or early life. For instance, prenatal PCB exposure affected male rat and mouse reproductive tract functioning later in life (Fiandanese et al., 2016; Hany et al., 1999), immune deficiencies or defects in a number of young mammals (Hertz-Picciotto et al., 2008), multigenerational and sex-specific growth effects in American kestrels (Falco sparverius; Fernie et al., 2003a, 2003b), and slow growth and mortality in juvenile snapping turtles (Chelydra serpentine) exposed to maternally offloaded PCBs (Eisenreich et al., 2009). However, it should be noted that contaminant vulnerability can change over the lifetime of an animal. For contaminants such as PCBs, which are metabolized in the liver and whose metabolites are often more toxic than the parent compounds, hepatic development could play an important role in when animals are most vulnerable. While contaminant effects are known to occur in utero, there is the potential for effects to be greater later in life after organ systems are fully developed; however, changes across multiple life stages will not be address in this thesis.

1.4 Organochlorine Effects in Elasmobranchs While effects of organochlorine contaminants on neonates and juveniles have been studied in a variety of taxa, there is a lack of information regarding organic contaminant effects in elasmobranchs. This gap is of concern as elasmobranchs are placed between jawless and bony fishes, and are part of a large number of chondrichthyan species that serve important ecosystem functions (Stevens et al., 2000). Elasmobranchs, primarily of the superorder Galeomorphii, are known for their top-predator roles within ecosystems. For example, tiger shark (Galeocerdo cuvier) presence or absence in marine communities alters the behavior of prey species, which in

6 turn has influenced primary production by sea grasses in small communities (Heithaus et al., 2008). In the tropical Pacific, the removal of large predatory sharks and tuna, by commercial fishing, altered the ecosystem so that mesopredators increased in abundance and dominance (Ward and Myers, 2005). Of the few studies directly examining effects of organic contaminant accumulation in elasmobranchs, none have investigated these effects at the neonatal or juvenile life stages. Gelsleichter et al. (2006) documented differences in reproductive steroids in adult Atlantic Stingray (Dasyatis sabina) taken from populations with varying levels of exposure to environmental organic contaminants. While some endocrine and immune disruption was observed in this study, the differences in reproductive output between the populations could have been due to factors other than organic contaminants. Manire (2002) investigated the potential for organic contaminants for causing reproductive failure in a population of Bonnethead Sharks (Sphryna tiburo) in Florida by analyzing a suite of reproductive characteristics from animals sampled from three different locations. There were few differences in the reproductive physiology of adults (i.e. hormone profiles, litter size, histology and immunohistochemistry of gonads) that could have been a cause for observed differences in birth rates. However, reproductive hormone concentrations were significantly lower in juvenile animals sampled from more anthropogenically-impacted sites. Other aspects of elasmobranch physiology have been implicated in reproductive and immune impairment. PCBs are known to negatively affect testicular steroidogenic pathways by interfering with hormone synthesis (Aydin and Erkan, 2017; Murugesan et al., 2005) as well as disrupt immune functioning (Martyniuk et al., 2016; Rice and Schlenk, 1995). Frantz (2014) sampled first-maturity male Round Stingrays (Urobatis halleri), administered a single dose of Arochlor 1254 (a common commercial PCB mixture), and found a decreased mRNA abundance of steroidogenic acute regulatory protein and 3β-hydroxysteroid dehydrogenase in the testes within 24 hours. In the same population of Round Stingrays, innate immune function in males was compromised by environmental PCB exposure compared to stingrays sampled from an offshore island in southern California (Sawyna et al., 2017).

1.5 Knowledge Gaps

Elasmobranchs are certainly capable of mounting a physiological response to contaminant exposure at a biochemical level (Alves et al., 2016; Lyons et al., 2014). However, most studies have focused on adults and reproduction. Little attention has been given to potential impacts in utero, despite the fact that maternal offloading of contaminants is wide-spread among elasmobranchs of varying matrotrophic strategies (Lyons et al., 2013; Lyons and Adams, 2015; Lyons and Lowe, 2015). In addition, other aspects of elasmobranch physiology are also likely to be affected by contaminant exposure, including growth and endocrine disruption (e.g. stress response). Current population management models do not take into account the differential costs of contaminant exposure that are moderated by ecology and physiology. For example, Salice et al. (2014) used matrix modeling to demonstrate that growth retardation in juvenile snapping turtles (Chelydra serpentina) from PCB exposure would have effects at the population level. As K- selected species, elasmobranch population models are sensitive to small changes in survivability and/or fitness of juveniles that are recruited into the population (Frisk et al., 2001). Therefore, examining physiological responses to contaminant exposure in young elasmobranchs is needed.

1.6 Round Stingray Ecology in Southern California

Elasmobranchs have been challenging to study due to their elusive nature and challenges of rearing them in captivity. Higher trophic status elasmobranchs that tend to accumulate the most contaminants are typically too large for captivity and laboratory experiments. The high vagility of many elasmobranchs also complicates the identification of an appropriate reference site for field studies where “control” animals with lower contaminant burdens can be obtained for comparisons. Thus, using smaller, easy to manage, elasmobranchs as models would prove useful for studies of elasmobranch toxicology that can then be extended to other species. The Round Stingray (Urobatis halleri) is locally abundant in southern California near sites of historic contamination (see below) and has many aspects of its physiology and ecology previously described (Hale et al., 2006; Mull et al., 2010; Vaudo and Lowe, 2006). In the early 1990s, California passed a nearshore gillnet ban to promote the protection of species commonly caught in this fishery, including White Seabass (Atractoscion nobilis), Giant Seabass (Stereolepis gigas) and Tope (Galeorhinus galeus; California State Proposition 132). While many of these species

8 have shown signs of recovery (Pondella and Allen, 2008), the removal of many of these predators probably contributes to their abundance in near-shore waters (Lowe et al., 2007). Round Stingrays reproduce with internal gestation and a form of matrotrophy called histotrophy, whereby females provide supplemental nutrition to embryos via uterine secretions (Wourms and Demski, 1993). Round stingrays have a fairly short gestation (3 - 4 months) compared to most elasmobranchs and embryos tend to grow in an exponential fashion until birth (Lyons and Lowe, 2013a). Gestation takes place in three phases (early, mid, late), according to the source of embryonic nutrients. During early gestation (approximately 1 month), embryos are completely dependent on their yolks for nutrition. During mid-gestation, embryos have exhausted their yolk and nutrition is supplemented by maternally-derived histotroph. Late gestation comprises the majority of gestation (~2-3 months) during which embryos are completely dependent on maternal uterine secretions. By parturition, embryos have increased by nearly 50x in mass compared to ovulated eggs and embryos are born at ~40% the disk width of their mothers. This is an exceptional maternal investment that is expected to be demanding for mothers also dealing with the metabolic challenges of a contaminated environment. Thus, round stingray are expected to show reproductive and metabolic deficits in response to exposure to environmental contamination.

1.7 Sources of Contamination at the Mainland Site

Southern California is a highly urbanized area, making anthropogenic influence on the marine environment particularly problematic. One major anthropogenic influence on the marine environment is the existence of a United States Environmental Protection Agency Superfund site at the Palos Verdes Peninsula. Historic chemical manufacturing of organic contaminants and subsequent disposal through the sewage system has led to very high levels of dichlorodiphenyltrichloroethane (DDT) in the area as well as PCBs. While sediment concentrations have declined after the cessation of chemical production in the early 1970s, since the 1990s concentrations have stabilized, which is attributed to bioturbation and the resuspension of contaminated sediment (Eganhouse et al., 2000; Eganhouse and Pontolillo, 2000). These contaminants continue to accumulate to high levels in local biota (Allen et al., 2004; Blasius and Goodmanlowe, 2008), resulting in recommended no fishing zones around the peninsula (Klasing et al., 2009).

Southern California is a relatively dry area with a Mediterranean climate such that the rainy season occurs during the winter and early spring with little to no rainfall during the summer and fall. Freshwater inputs into southern California wetlands are limited due to low rainfall; thus, the large majority of wetlands are saltmarshes with salinities nearly identical to seawater. In addition, outfalls for many of southern California’s rivers have been modified to direct flow through concrete channels as a result of heavy urbanization. Similarly, storm water runoff is directed to the ocean through pipes. This results in freshwater flowing directly into the ocean, rather than through wetlands. In addition, due to southern California’s large urban population, there are four major wastewater treatment plants along the coast that discharge > 100 million gallons of treated water into the marine environment every day (Steinberger and Stein, 2003); however, that water is discharged through subsurface pipes that release water ~ 8 km offshore in water that is > 200 m deep1. Thus, despite the fact that southern California has one of the largest DDT signatures in the world, DDT contributes little to total legacy contaminant burden in Round Stingrays. Rather, PCBs comprise ~75% or greater to total contaminant loads for those contaminants where standards are available for screening (Lyons et al., 2014). Concentrations in Round Stingray livers typically exceed the World Health Organization’s daily allowable PCB intake of 6 µg/kg per day (American Academy of Pediatrics Committee on Environmental Health., 2003). The lack of a strong DDT signal in Round Stingrays likely reflects their ecology and preference for being nearshore in areas that are calm, warm, and relatively shallow, which would prevent them from spending much time in the deeper waters (> 100 m) on the Palos Verdes Shelf where urban outflows are directed. The mainland site also has low exposure to environmental pharmaceuticals for the same reason. Treated sewage water is discharged offshore away from areas stingrays prefer. Wastewater pharmaceuticals do have effects in biota that use habitat near the outfall pipes in southern California (Vidal-Dorsch et al., 2013). Round Stingrays have been tested for pharmaceutical contamination, with low frequency of detections for measured compounds and low concentrations of those that have been above the limit of detection (e.g. diphenhydramine, an antihistamine; Lyons, unpublished data). Thus, with DDT and pharmaceutical contaminants as

1 https://www.ocsd.com/residents/current-construction/outfall-land-sections-oobs-piping-rehabilitation-project-in-the- city-of-huntington-beach

10 minimal environmental burdens for this population of Round Stingrays, the mainland site represents a notably specific PCB-only contamination burden that is rarely found in field-based research.

1.8 Comparisons with the Offshore Site

Offshore islands occupy the same geographical area, while controlling for anthropogenic influence. Santa Catalina Island is located approximately 35 km from the California mainland. A deep ravine in the San Pedro Channel separates the island from the mainland, making it difficult for benthic-oriented species, without a pelagic phase as part of their development, or species with limited home ranges to cross from the mainland to the island. The California Countercurrent (i.e. Davidson Current) flows from Mexico northwards towards the northern Channel Islands where it meets the south-flowing California current. This has the effect of limiting transport of contaminants from the Palos Verdes Shelf to the Island. Santa Catalina Island had a resident population of 4096 in 2010 with 91% of residents in the town of Avalon2, which is removed from the sampling site. Previous research has already demonstrated that PCB contamination burden is lower in Round stingrays sampled from this site at Santa Catalina Island is lower than at the mainland site (Lyons et al., 2014; Sawyna et al., 2017). The divide between the island and the mainland is thought to prohibit stingray movement between the sites, and prevent gene flow (Plank et al., 2010). A benefit of this separation means that there is a high degree of confidence that stingrays sampled from each location are representative of that area, reducing concern about stingrays acquiring contaminants at the mainland and swimming over to the island. The genetic markers used to determine separation between sites operate on drift, as microsatellites are chosen based on non-coding regions of DNA. Therefore, the study that detected the genetic differences did not provide any relevant measures of physiological differentiation. Although it remains possible that the selection pressure of contaminants has selected for genetic change in the mainland populations, the selection interval has been short. The peak of organic contaminant production occurred in 1970, and this study was conducted only 44 years later (2014), which is equivalent to ten generations. Furthermore, based on mitochondrial DNA,

2 https://en.wikipedia.org/wiki/Santa_Catalina_Island_(California)

11 elasmobranch evolution occurs at a slower rate than mammalian evolution (Martin et al., 1992). Therefore, genetic divergence is not expected to play a major role in physiological responses to contaminant exposure. Another environmental difference that does not differ between sites is salinity. Like the majority of southern California wetlands, the mainland site is comprised of saltmarshes that are fully tidally inundated and receive little freshwater input. Thus, the mainland site where stingrays in this study were sampled has salinities nearly identical to pure seawater (Merkel and Henderson, 2014). Santa Catalina Island has no rivers and, thus, the only freshwater input comes with the rainy season making both sites equivalent to pure seawater. The primary environmental factor that differed between the sites, other than PCB contamination, was water temperature (Crear, 2015; M. Shane, pers comm). The structural shelter afforded by berms and shallow channels at the mainland sampling site resulted in higher water temperatures during the sampling interval compared to the site at the island. Temperatures obtained from Crear (2015) were within the marsh complex where females from the mainland site were sampled, and temperatures taken in one pond were assumed to be similar among the complex of connected ponds. At the island sampling site, temperatures were taken from fish pens located within the bay where females were sampled. Although pens were located slightly closer to the mouth of the bay than the exact sampling location (~ 100 m), since the bay is fully tidal, temperatures measured at the pens were assumed to be similar to the sampling location. Higher temperature was predicted to enhance reproductive success by shortening gestation (Hight and Lowe, 2007) and is, therefore, considered as an important environmental factor in this research. Enzyme activities are also influenced by temperature. Thus, laboratory conditions were standardized to room temperature for all measures of enzyme activities between sites, and temperature was considered as a variable in interpreting all findings in this thesis. The current research, therefore, will use this two-site comparison between mainland and Catalina Island. With geographic proximity and equivalent to pure seawater salinity, the sites differed only in water temperature. With genetic isolation between the two populations based on deep-water separation, there was no expectation of movement of individuals between sites. In terms of contaminant exposures, the PCB exposure difference was previously established, and both DDT contamination and pharmaceutical contamination were previously established as

12 minimal at the mainland site. Thus, water temperature was considered as a possible explanation for all findings, and other differences were cautiously attributed to PCB exposure.

1.9 Research Objectives

Using two populations of Round Stingrays at the relatively pristine site of Catalina Island and the anthropogenically-impacted mainland site, the overarching objective of this thesis is to assess the lifetime physiological impact of organochlorine contamination. Specifically, I hypothesize that: 1. Adverse effects of organochlorine contaminant exposure will impact adult females during the energetically demanding pregnancy phase of the lifespan. 2. Adult males will have lower energy demands relative to females undergoing gestation, reducing impacts from organochlorine contaminant exposure. 3. Maternal organochlorine contaminant exposures, compounded by maternal offloading of lipophilic contaminants to embryos, will adversely impact embryonic development, with lifelong consequences. To test these hypotheses, this thesis will examine diverse measures of physiological homeostasis in Round Stingray at the reference (Catalina Island, CA) and anthropogenically- impacted (Mainland, CA) sites. The following objectives will be explored: assessment of effects of exposure on pregnant female stingrays captured throughout their reproductive cycle for their energetic responses to pregnancy (Chapter 2), embryonic contributions to homeostasis in their embryonic environment (i.e. histotroph; Chapter 3), the ability of adult Round Stingrays (both males and females) to mount an adaptive response to an externally-applied acute stressor (Chapter 4), the ability of pregnant females to osmoregulate in their marine environment (Chapter 5), the ability of adult male and female stingrays along with embryos across ontogeny to support metabolic processes essential to detoxification and reproduction (Chapter 6), and the cumulative effects of contaminant exposure on Round Stingray maintenance of homeostasis and successful reproduction (General Discussion).

Chapter 2: Physiological Consequences of Legacy PCB Contamination in an Elasmobranch with Matrotrophic Histotrophy, the Round Stingray (Urobatis halleri): Impaired Intrauterine Development

Anthropogenic chemical exposure can result in overall reductions in reproductive success. Using the Round Stingray (Urobatis halleri) as an elasmobranch model with internal gestation, we measured female fecundity and embryo growth from post-ovulation to near parturition to test the hypothesis that environmental PCB contamination would impair reproductive success. Two sites were sampled from southern California: the mainland site was exposed to legacy PCB contamination (with low exposure to other anthropogenic contaminants), and the offshore reference site at Catalina Island was a separate population with low anthropogenic influence. Reference females ovulated later than contaminant-exposed females; an effect likely mediated by the colder water temperatures at Catalina. Despite exposure to warmer temperatures, PCB- exposed embryos weighed less at each stage of development than reference embryos, but accumulated proportionately more liver lipids over development. Furthermore, environmental contamination had sex-specific effects, negatively affecting males more than females. Litter mass variability significantly increased over development, suggesting the existence of intrauterine competition as well as the potential for parent-offspring resource conflict. This is the first study to demonstrate a negative effect of contaminant exposure on elasmobranch embryo growth, with probable fitness costs later in life.

2.1 Introduction Polychlorinated biphenyls (PCBs) are resistant to degradation (Chandra and Chaudhary, 2013) and, therefore, represent one class of persistent chemicals that continues to adversely impact wildlife health although they are no longer produced (Eisler and Belisle, 1996). PCBs, like other organic contaminants, have the ability to impair hormone signaling (Crisp et al., 1998) and disrupt normal metabolic processes (Swedenborg et al., 2009), among other negative effects. PCBs also affect reproduction at both the maternal (Hose et al., 1989; Johnson et al., 1988; Reijnders, 1986) and embryonic level (Bergeron et al., 1994; Goldey et al., 1995; Hoffman et al., 1986).

Elasmobranchs (sharks, skates and rays) exhibit traditional K-selected ecological traits (Musick, 1999), as they mature late and produce few young over the course of their lifetime. Many elasmobranchs are mid to top level predators with long lifespans, giving them a propensity to accumulate high levels of contaminants through the process of bioaccumulation (contaminant accumulation over time) and biomagnification (contaminant concentration with increasing trophic level) (Fisk et al., 2002; Lyons et al., 2013). Despite these characteristics, little work has examined reproductive effects in elasmobranchs with respect to organochlorine contaminants. Circulating hormone levels and fecundity were explored in two populations of Bonnethead Sharks (Sphyrna tiburo) occupying areas with differential contaminant exposure (Manire, 2002). However, this study found no significant connection between reproductive indices and contaminant exposure. Focusing only on female fecundity may cause us to overlook other potential effects resulting from contaminant exposure. Examining effects across the life span is important to pinpoint which stages are the most susceptible to exposure. For many vertebrates, the youngest life stage is the most sensitive to contaminants (Elonen et al., 1998; Guillette et al., 1994; Örn et al., 1998), and response to exposure during this time may be more pronounced than similar levels of exposure in adults that have already completed development. Contaminants may also have other impacts on fecundity outside of reducing offspring number, whereby the quality of offspring is reduced. For instance, juvenile Common Sole (Solea solea) fish collected from anthropogenically-impacted sites showed reduced growth and lipid stores compared to fish from less impacted sites (Amara et al., 2007). Therefore, incorporating several indicators of reproductive function is necessary to assess the effects contaminants could have on elasmobranch reproductive success. Characteristics of elasmobranch biology, such as high vagility, slow development, and ethical sampling concerns (Stevens et al., 2000) have impeded investigations into contaminant effects in this taxon. Therefore, utilizing a more accessible and abundant species as an elasmobranch model may help to overcome traditional challenges in studying contaminant effects in elasmobranchs when using field studies. The Round Stingray (Urobatis halleri) is a common, coastal species found throughout southern California in fairly high numbers during the spring and summer (Hoisington and Lowe, 2005), with a well characterized reproductive cycle (Babel, 1967; Mull et al., 2010). As maternal contaminant offloading has been demonstrated in this species

(Lyons and Lowe, 2013a), and two populations of Round Stingrays with differential exposure to primarily PCB contaminants identified (Lyons et al., 2014; Sawyna et al., 2017), the current study was designed to test the hypothesis that stingrays sampled from the impacted site would have impaired reproductive success (measures of fecundity, embryo growth rate, and offspring quality) compared to stingrays sampled from our reference site.

2.2 Methods 2.2.1 Study sites. Stingrays were sampled from two locations within southern California, one from the mainland and one from an offshore island. Mainland California was considered our contaminant- exposed site and Santa Catalina Island, located 35 km offshore and separated by a deep channel (> 700m in depth) from the mainland, was considered our reference site based on previous knowledge of historical release of contaminants in the two respective areas (Eganhouse et al., 2000; Eganhouse and Pontolillo, 2000). Difference in environmental release has resulted in stingrays from the mainland population having significantly higher levels of accumulated organic contaminants than the island population (Lyons et al., 2014; Sawyna et al., 2017). Despite the preponderance of DDT in the Southern California Bight, PCBs comprise the majority (> 75%) of organochlorine contaminants in stingrays (Lyons et al., 2014), which is attributed to their ecology and preference for nearshore habitats that separate them from deeper, more diversely contaminated sediment where historic and current outfall pipes release wastewater (~8 km offshore and >200m deep). Stingrays’ nearshore preference also results in low exposure to treated wastewater, with most screened pharmaceuticals measured in hepatic tissue below detection limits (Lyons, unpublished data). Both sites are completely inundated twice daily with full seawater, and experience little fresh water run-off from land both because the area has a long dry season that encompasses the reproductive season, and because urban runoff from the mainland is also collected and diverted offshore. Thus, the two full seawater marshes located in the Seal Beach National Wildlife Refuge (SBNWR; 33.731N, 118.064W) and from the back bay of Catalina Harbor (33.434N, 118.503W; Figure 2.1) differ primarily in their exposure to legacy PCB contaminants.

Figure 2.1. Map of sampling sites Stingrays were sampled (black dots) from two areas in southern California. Seal Beach (Mainland California) represented our contaminated site and Santa Catalina Island, located ~35 km offshore, represented our reference site. Star indicates the location of the Palos Verdes US Environmental Protection Agency superfund site and relative depth is depicted in shades of grey, with darker colors corresponding to deeper depths. Arrows indicate direction of the California Countercurrent, reducing the environmental transport of contaminants from the mainland to the island.

2.2.2 Sampling. Female Round Stingrays were captured monthly during the summer of 2014. Males were rarely captured (n = 7 and 14 for the two sites) and, therefore, excluded from all analyses. Upon capture via hook and line, stingrays were subjected to one of two treatments (“baseline” and “stressed”) in order to also examine the effect contaminant exposure had on the acute stress response (Chapter 4). However, given the short duration of our stressor (~15 mins), I did not include stress as a factor when assessing contaminant effects on stingray reproductive output as the short duration would not have an immediate impact on the reproductive variables measured. Stingrays were euthanized with an overdose of tricaine methanesulfonate (MS-222) in accordance with animal care protocols (University of Calgary Animal Care Protocol #14-0016) and disk width (size across the body) was recorded prior to tissue sampling. For pregnant females, any embryos were excised after euthanasia, wrapped in foil, and frozen for later dissection. 2.2.3 Data analysis. R statistical package (V 2.13.0) (Team, 2011) was used for all analyses and α was set to 0.05. 2.2.3.1 Environmental factors. As water temperature influences reproductive timing in Round Stingrays (Mull et al., 2010), and is suggested to benefit embryos by promoting growth (Hight and Lowe, 2007; Jirik and Lowe, 2012), it was important to determine if sites were comparable. Hourly data logger temperatures were collected in the context of a different study in the SBNWR from mid-April to mid-August Crear (2015) and from Hubbs SeaWorld Research Institute from fish pens located within Catalina Harbor from April to mid-July (M. Shane, pers. comm.). Temperatures from these studies were taken in close proximity to sampling sites and were assumed to be reflective of temperatures experienced by female stingrays. Daily averages were compared between sites using a paired t-test. 2.2.3.2 PCB quantification. Although previous studies have documented significant differences in PCB concentrations in the livers of reference and contaminant-exposed stingrays (Lyons et al., 2014; Sawyna et al., 2017), this observation was confirmed by measuring PCB contaminants in a subset of female liver samples (n = 16, eight from each location). Fifty-three PCB congeners were analyzed following the methods of Lyons et al. (2014). As expected, hepatic PCB concentrations averaged

18 significantly (4-fold) higher in contaminant-exposed females than reference females (Welch’s t- test, t8= 3.09, p = 0.015; Table 2.1).

Table 2.1. Organochlorine concentrations in a subset of stingrays Organochlorine contaminants were measured in a subset of female stingrays from the contaminant-exposed site (i.e. mainland California) and our reference site (Catalina Island). Contaminants were grouped into two types and represent the sum of congeners detected in that group: polychlorinated biphenyls (PCBs) and pesticides (e.g. dichlorodiphenyltrichloroethane and its metabolites, chlordanes). Within-site mean ± standard deviation is given as well as the contribution of PCBs to the total contaminant load. All contaminants are presented on a lipid weight basis (ng/g lw). Stingray # PCBs Pesticides % PCBs Reference site C1 496 70 88 C2 403 11 97 C10 122 31 80 C15 274 88 76 C18 386 19 95 C19 1,762 117 94 C20 3,211 370 90 C25 414 112 79 C28 285 96 75 Mean ± SD 817 ± 1,019 102 ± 108 86 ± 8.8 Exposed site M1 2,373 307 89 M5 826 132 86 M12 1,250 337 79 M15 1,711 206 89 M16 1,931 673 74 M20 715 199 78 M28 10,915 1,633 87 M29 3,128 487 87 Mean ± SD 2,856 ± 3,353 497 ± 491 84 ± 5.7

2.2.3.3 Fecundity. As with other elasmobranchs, stingrays continually grow throughout life and fecundity is related to female size (Hussey et al., 2010). Therefore, the number of embryos or eggs within a litter was linearly regressed against female disk width (DW) and slopes of the lines between the two sites were compared using an analysis of covariance (ANCOVA). To determine how litter biomass, accounting for female size (DW), changed over pregnancy I performed linear regressions against Julian sampling date, followed by an ANCOVA to examine differences between sites. 2.2.3.4 Developmental stage. To determine if there were embryonic growth differences between sites, I compared mean litter mass based on a developmental marker (“clasper days”) relative to a baseline measurement (Julian sampling date). Claspers are the male copulatory organs used by elasmobranchs, which develop as an extension of the distal ends of the paired pelvic fins. Clasper days were determined as the positive or negative interval relative to the first date on which male or female embryos for each site could be sexed by eye. Therefore, I used sex differentiation as a confident anchor for developmental stage comparison between sites and synchronization. Boltzmann sigmoidal curves were fitted to the data based on their high r2 value relative to other models (e.g. linear regression, second-order polynomial). 2.2.3.5 Sex-specific effects. Contaminants are known to have sex-specific effects in embryos exposed to maternally transferred contaminants (Kaya et al., 2002), as PCBs can exhibit either estrogenic or antiestrogenic properties depending on examined congeners (Jansen et al., 1993). To test whether contaminants differentially affected male embryos, I compared the relative mass (i.e. individual mass/mean litter mass) of the largest embryo in a litter by sex and site using student’s t-tests and Wilcoxon U-tests, with the a priori assumption that a sex effect would exist. Largest embryos were further compared by dividing litters into three groups based on their composition of males and females (i.e. female skewed, even sex ratio, male skewed), assuming that more females might represent a more “estrogenic” environment while more males might represent a more “androgenic” environment that could potentially influence growth. Deviations from the presumed 1:1 sex ratio were examined within sites using χ-square tests.

2.2.3.6 Embryo quality. Elasmobranchs use their livers as their primary energy storage organ (Rossouw, 1987). Therefore, liver mass and lipid content were used as proxies to assess embryo quality, especially since neonate elasmobranchs rely heavily on hepatic resources after birth (Hussey et al., 2010). Given the strong association between embryo mass and liver mass, embryos were compared on a total mass basis (rather than developmental basis, i.e. clasper days), to control for the established growth differences between sites. Liver mass was obtained from embryo dissections and lipids were extracted from one embryo per litter (Folch et al., 1957). Regressions were fitted for liver mass, liver lipid content, or total liver lipids against mean litter mass by site and sex. I further explored site-related effects in total hepatic lipids by expressing differences between sites by sex as a proportion of embryo mass over development. 2.2.3.7 Intrauterine growth variability. As litter number tends to increase with female size, I might expect competition to increase as embryos compete for the same, limited resource, which could result in unequal distributions leading to differences in size. I tested our hypothesis that increases in litter number positively related to litter mass variability (i.e. standard deviation normalized to mother’s disk width) over development using linear regressions followed by an ANCOVA with litter size as an ordinal covariate. Increased variation may result from greater differences between the largest and smallest embryos in a litter as litter size increases; however, disparity could arise from paternal differences as well (Lyons et al., 2017). Within-litter mass disparity with respect to litter number was explored via linear regression using embryo mass as a proportion of the mean litter mass.

2.3 Results 2.3.1 Timing of ovulation. At the first sampling event, 100% of contaminant-exposed females (n = 5) had visible embryos compared to 28% of reference females (n = 7), suggesting that populations experienced significant differences in ovulation timing. Back calculation based on the latency of reference females to reach 100% indicated that the lag to ovulation was approximately 30 days. The significantly warmer water temperatures at the SBNWR compared to Catalina Harbor is likely responsible for earlier ovulation in contaminant-exposed females (paired t-test, t80 = 26.7, p < 0.0001).

2.3.2 Fecundity. Female fecundity was comparable between sites. Size range (disk width) and number of pregnant females sampled at each site over the study were approximately equal for reference (19.5 – 23.8 cm, median = 22 cm, n = 27) and contaminant-exposed (19.7 – 24 cm, median = 21.2 cm, n = 28) samples. Median litter size was three embryos. As expected, litter number increased significantly with female size at both the reference (p = 0.003, r2 = 0.38) and contaminant- exposed populations (p = 0.007, r2 = 0.36). Site did not affect the relationship between female size and litter number (ANCOVA, p = 0.48; Figure 2.2). Litter biomass significantly increased over pregnancy, with no difference in rate of increase between sites (ANCOVA, p > 0.2; Figure 2.2). The proportion of total litter mass relative to their mother’s mass (dressed mass + liver mass) increased linearly with pregnancy. Dressed mass (body mass without internal viscera) was used as it removes variability added from stomach or intestine fullness or variability contributed by other organ masses. By the end of pregnancy, embryos comprised 15-20% of their mother’s dressed mass, demonstrating the high resource investment of female Round Stingrays into offspring production. 2.3.3 Developmental stage. Despite both the significantly warmer water temperatures at the SBNWR compared to Catalina Harbor, and the consequently earlier ovulation by females sampled at the SBNWR, aligning litters to a developmental marker (clasper days) revealed clear site differences. Reference litters were heavier than their contaminant-exposed counterparts at all developmental stages (Figure 2.3), and bigger, particularly during early to mid development (Figure 2.3). This supports our hypothesis that contaminant exposure would have a negative effect on embryo growth.

A 6 Reference Exposed

4 # Embryos 2

0 20 22 24 Female Disk Width (cm)

B Reference Exposed 6

2 Mean Litter Mass / Maternal DW 0 Jun 1st July 1st Aug 1st Sept 1st Julian Sampling Date

Figure 2.2. Female reproductive success Fecundity was similar between females sampled from reference site (circles, grey line) and our contaminant-exposed site (triangles, black line). Litter size significantly increased with female size and was not different between sites (A). Mean litter mass (corrected for mother’s size, DW) significantly increased across pregnancy and rate of increase was not different between sites (B).

Figure 2.3. Embryo growth can be aligned by a developmental marker, clasper day Embryo growth (mean litter mass) was compared between reference (circles) and contaminant- exposed (triangles) litters two different metrics. Julian sampling date (A) was used as our baseline “reference”, which did not take into account reproductive asynchrony between populations or the effect of differences in temperature. Temperature was significantly higher in the SBNWR (exposed site) than at the island (reference site) based on data collected from water loggers at the two sites during the time of sampling (B). When litters were aligned based on a developmental marker, such as the timing of sexual differentiation by the presence/absence of visible claspers (i.e. clasper days), reference litters were found to be both heavier in terms of mean litter mass (C) and larger in terms of disk width (D) compared to their contaminant-exposed counterparts. Boltzmann sigmoidal curves were fitted to the data to visualize growth patterns.

2.3.4 Sex-specific effects. Site also had an interaction with sex for influencing embryo growth. The largest reference male embryos in a litter were significantly heavier than both the largest reference females (t18=

2.9, p = 0.005) and contaminant-exposed embryos regardless of sex (M-M: t16 = 3.2, p = 0.003; M-F: W = 85, p = 0.036; Figure 2.4). Largest females showed no difference between sites (W = 65, p = 0.5). For reference litters, the frequency of whether a male or a female would be the largest embryo in a litter was even (i.e. male:female ratio of 1:1). While sex ratios for the largest embryo in contaminant-exposed litters was skewed towards females (1:1.38), it did not significantly deviate from the presumed 1:1 ratio (χ-square, p = 0.5). When comparing largest embryos based on litter sex composition, I found no differences in relative mass for litters with proportionally more females or litters with an even sex composition (p = 0.23 and p = 0.15, respectively; Figure 2.4). However, for male-skewed litters, I found that reference litters were significantly heavier than contaminant-exposed litters (t10 = 5.09, p = 0.0002). 2.3.5 Embryo quality. The importance of the liver to elasmobranch development was highlighted by the exponential and linear increase in liver mass and lipid content, respectively, over development with hepatosomatic indices (ratio of liver mass to total mass) ranging from 1.55 to 5.60% in reference and 1.57 to 6.08% in contaminant-exposed embryos. For the largest embryos in a litter, sex-related differences in hepatic metrics became apparent late in development when embryos were aligned by total mass (Figure 2.5). No differences were found between sites for the largest females in a litter for either liver mass or total hepatic lipid content with respect to size. On the other hand, I found significant differences between males. Liver mass was significantly lighter in reference males and these differences were further exacerbated when comparing total hepatic lipids. By the end of development, contaminant-exposed males were estimated to have ~50% greater lipid content than reference males (Figure 2.5), while differences between females decreased with development to the point where females were equivalent as they approached near parturition size. Therefore, I found that embryo quality based on hepatic metrics had a strong sex-related effect.

Figure 2.4. Sex and site differences in mass of the largest embryo. Relative mass of largest embryos among litters was compared between males and females by site. (A) Reference male embryos (n = 10) were significantly heavier, relative to their littermates, compared to both reference females (n = 10) and contaminant-exposed males and females (n = 8 and 11, respectively). No sex effect was found in contaminant-exposed litters with dominant male embryos having the same relative proportionate mass as dominant females. Crosses denote group means and different letters indicate significant differences among groups. (B) To account for any potential effects related to sex, relative mass of the largest embryos were also compared by dividing litters based on sex composition. Male skewed litters showed a significant site difference with reference males being proportionally heavier than contaminant-exposed males. Different letters represent significant differences within pairs.

Figure 2.5. Hepatic lipids by sex and site. Embryo quality was determined by comparing embryos on a mass basis (total mass) rather than a developmental basis (clasper days) as I was interested in quality differences for embryos of the same size. (A) No differences were found between females (light grey symbols) by site; however, contaminant-exposed males (dark grey triangles, solid line) had larger livers compared to reference males (dark grey circles, dashed line). (B) Differences between males were more exaggerated when comparing total hepatic lipids, while females showed no differences. (C) The relative difference in total lipid content of male and female embryos from reference and contaminant-exposed sites, expressed as a percent of reference embryo theoretical mass (based on B curves), was calculated and plotted.

2.3.6 Intrauterine growth variability. As litter size increases, I might expect there to be a tradeoff between individual embryo mass and number of embryos (Elgar, 1990). The largest embryo a female produced was not affected by litter size, as the largest embryo from litters of two and four were equal in mass (~43 g; Figure 2.6). However, production of equally heavy embryos from different litter sizes could come at the expense of the largest embryo’s littermates. As litter number increases, I hypothesized that intrauterine competition would also increase, leading to greater variability within litters. I observed that litter mass variability (i.e. standard deviation normalized for mother’s size) increased as development progressed (r2 = 0.48, p < 0.001; all litter sizes pooled), with the potential for litter size to exacerbate these effects (Figure 2.7). In support of our hypothesis that increased litter variability is manifested through differences between the largest and smallest embryo, I found that the relative mass of the largest embryos to the litter mean positively increased with litter size (r2 = 0.22, p = 0.009), while the relative mass of the smallest embryo weakly decreased (p = 0.06; Figure 2.7).

Figure 2.6. Mass variability within litters. The largest embryo by mass from each litter was compared across litters of varying sizes over the course of development. The horizontal line denotes the maximum mass measured in this study (43 g), which was observed in litters of two and four.

Figure 2.7. Within-litter competition for resources. Variability within litters significantly increased as development progressed (A) for all litters combined (solid line), litters of four (square), three (triangle) and two (circle), suggesting the presence of intrauterine competition. (B) Individual mass relative to litter mean significantly increased for the largest embryo in a litter (squares, solid line) and weakly decreased for the smallest embryo (diamonds, dashed line), indicating that dominant embryos are able to better monopolize resources. Horizontal, dashed line represents the null hypothesis with mass of the embryo equal to mean litter mass.

2.4 Discussion In general, maternal fecundity was not affected by site, whereas embryonic indices (i.e. growth and quality) showed significant impairments at the contaminated site. This pattern was consistent with contaminant studies in other fishes noting effects to be ontogenetic-specific (Zhou et al., 2001), with embryonic life stages being more sensitive to contaminant exposure than others (Elonen et al., 1998; Örn et al., 1998). While water temperatures were higher at the contaminated site, which was predicted to have a positive effect, indices of embryo growth and quality were lower than reference site embryos. Despite the genetic divergence between our two populations (Plank et al., 2010), we do not believe that genetic difference was unlikely to have played a major role in physiological responses of stingrays to contaminant exposure. First, non-physiologically relevant portions of DNA were used to assess population structure (microsatellites), which did not provide any relevant measures of physiological differentiation. Second, the interval of time between contaminant production peak and sampling has been short. The peak of organic contaminant production occurred in 1970, and this study was conducted only 44 years later (2014), which is equivalent to ten generations. Based on mitochondrial DNA, elasmobranch evolution occurs at a slower rate than mammalian evolution (Martin et al., 1992). Therefore, genetic divergence was not expected to play a major role in physiological responses to contaminant exposure. Considering we accounted for differences in water temperature, the second most important variable that differed between sites, we cautiously attribute our findings to effects related to PCB exposure. Thus, the absence of clear fecundity effects in females cannot rule out the potential for adverse impacts of pollutants at early stages of development. 2.4.1 Fecundity. Adult stingrays did not experience a reduction in litter number when sampled from an anthropogenically-impacted site compared to a relatively pristine site with different temperatures. Therefore, in terms of individual fitness based strictly on offspring number, contaminants and temperature did not affect adult females. However, with respect to reproductive timing, temperature rather than contaminants played a more influential role. As previously demonstrated (Mull et al., 2010), water temperature played a key role in female Round Stingray reproductive timing, resulting in later ovulation in reference females that experienced colder temperatures compared to their mainland counterparts sampled at the SBNWR. Open ocean temperatures were not different between sites (Figure 2.8), and our study reinforces the influence that availability of

31 thermal refuge habitat has on stingray reproductive timetables (Jirik and Lowe, 2012; Figure 2.8), and likely has implications for other elasmobranchs (Hight and Lowe, 2007). 2.4.2 Developmental stage. In line with other taxa, negative effects on embryonic growth and differences between sites were most apparent in late-term embryos. As is also common when temperature is an important predictor of reproductive timelines, asynchrony in reproductive events between populations required alignment with a developmental marker in order to reveal differences. Clasper days represented our best way of aligning litters, although we acknowledge that we do not know the effect that temperature or degree of maternal investment plays on the first appearance of claspers. Clasper days indicated that growth was impaired in contaminant-exposed litters compared to reference litters. Reference embryos were larger by mass than contaminant- exposed embryos both before and after clasper appearance. Since the SBNWR had significantly warmer water temperatures, an environmental factor purported to accelerate embryo growth and shorten pregnancy (Hight and Lowe, 2007; Jirik and Lowe, 2012), I expected contaminant-exposed embryos to be heavier and more developmentally advanced than reference embryos when compared based on sampling date alone. However, I found mean litter mass to be comparable between sites. Therefore, if warmer temperatures do promote embryo growth, this benefit appears to be inhibited by the negative effect of contaminants on growth. Taken together, despite the fact that contaminant-exposed embryos had a “head start” (i.e. earlier ovulation) and warmer water temperatures (increased metabolism), reference embryos were able to catch up in size to contaminant-exposed embryos by mid- development (i.e. reduced reliance on external yolk sac). Our results indicate that contaminants do have a negative effect on embryo growth and corroborate other studies demonstrating similar effects (Adams et al., 1992; Hoffman et al., 1986). 2.4.3 Sex-specific effects I also found evidence that the negative effect of contaminant exposure on embryo growth and quality is sex-specific. Accounting for litter variability (see below) by only using the largest embryo from a litter, the pattern in our reference populations suggests that males in a litter should be significantly heavier than females. However, in our contaminant-exposed litters, I found that males were not different than females and were significantly lighter than reference males. PCBs, the primary class of legacy organochlorine contaminants measured in Round Stingrays, are

32 known to have estrogenic as well as antiestrogenic endocrine disrupting effects depending on model and factors tested (Hany et al., 1999; Kaya et al., 2002). Thus, the absence of site differences between females, but with smaller males at the contaminated site, support the hypothesis that contaminants have sex-specific growth impacts in stingrays. Considering that male-male competition is high in Mainland Round Stingrays (Lyons et al., 2017), males that are smaller in early life could be disadvantaged in their later reproductive success, impairing lifetime fitness; however, this would need to be explored further by linking male size to reproductive success.

Figure 2.8. Water temperatures by site. Water temperatures were collected via two methods from within estuaries (data loggers) and just outside of the estuary mouths (satellite sea surface temperatures), respectively. Open ocean sea surface water temperatures outside estuary mouths were retrieved from the online NOAA ERDDAP database (Aqua MODIS). NOAA satellite temperatures showed no significant difference in open ocean temperature between sites (A); however, temperatures within the SBNWR (i.e. Mainland logger data) were significantly warmer than NOAA temperature readings (B). On the other hand, temperatures were not different NOAA satellites and data loggers, likely due to the open, physical structure of this embayment with more ocean influence.

2.4.4 Embryo quality. Since contaminant detoxification represents an additional energetic cost (Beyers et al., 1999), and the liver is the main energy storage organ of elasmobranchs, I predicted that embryos from the impacted site would have lower hepatic quality relative to reference embryos. However, I found that, similar to total mass, hepatic quality differences between sites were sex-specific. The absence of differences between females, but significant differences between males, suggests that contaminants are having sex-specific effects on embryo liver metabolism in addition to overall growth effects. Contaminant-exposed male embryos had both larger livers and higher total lipid content when equally sized male reference embryos were compared based on mass, rather than on development (i.e. clasper days). The lower quantity of lipids and smaller livers (mass-basis), but larger embryo size (developmental-basis), could indicate that reference male embryos are more efficient at converting maternal resources (i.e. histotroph) into extra-hepatic somatic growth. The liver is the major lipid storage organ in elasmobranchs, and has a propensity to accumulate organochlorine contaminants, making it particularly susceptible to contaminant effects. In teleosts, heavy metal contaminant exposure disrupts proper lipid metabolism and seasonal cycling (Levesque et al., 2002). Similar types of effects could be occurring in embryonic rays and could have fitness implications later in life if neonatal stingrays are unable to properly utilize their lipid stores. Interestingly, our observations of sex-related effects mirror those found in other studies on adult stingrays (Lyons et al., 2014) where males showed significant differences in their biochemical responses to contaminant exposure between sites but females did not. Therefore, future studies should examine whether these sex-related differences persist throughout an animal’s life and, if so, what the implications it has for male versus female fitness are. 2.4.5 Intrauterine competition. I found evidence that intrauterine variability increases over development and is potentially exacerbated by litter number, suggesting the presence of intrauterine competition. As embryos grow, the ratio of uterine space (i.e. histotroph availability) compared to embryo volume decreases, reducing resource availability, thus leading to increased resource competition as development proceeds. If embryos all had the same access to resources, litter mass variation would be expected to be relatively low and not change. However, I observed litter mass variability to significantly increase with development, supporting the intrauterine competition

35 hypothesis, which has also been documented in swine Sus scrofa (Baker et al., 1958). I also saw indications that larger litter number might also intensify competition, as more littermates would reduce histotroph availability. Although differences were not significant, litters of four had greater than average increases in variability with development, followed by a stepwise decrease in litters of three and two. Based on mass of the largest embryo in a litter, I found no indications that mothers experience a tradeoff between litter size and “biggest” offspring mass as demonstrated in some mammalian species (Mappes et al., 2004). Elasmobranch females experience indeterminate growth and their larger physical size may allow them to accommodate larger litters with reduced costs to offspring mass. However, the ability of females to maintain large embryos of comparable mass with increasing litter sizes could be due to a tradeoff between litter size and mass variability. In this study, the largest embryos in a litter were able to obtain significantly higher relative masses as litter size increased, while the opposite was true for the smallest embryos. As competition intensifies with increasing litter number, embryos that are better able to monopolize resources become more visible when resources are more limited in larger litters. Alternatively, the smallest embryos in a litter might become worse at obtaining resources with larger litter sizes, which further contributes to rising litter variability with size. The fact that intrauterine competition exists suggest that resource conflict may be present between mothers and embryos (Trivers, 1974). Mechanisms by which embryos might manipulate their mothers to maximize resource acquisition are unknown; however, elasmobranchs represent unique models to explore parent-offspring conflict (Crespi and Semeniuk, 2004) given the range in matrotrophic strategies utilized. 2.4.6 Conclusions. Besides differences in PCB concentrations, water temperature was the next variable to significantly differ between sites. By accounting for this variable, we attribute our results to PCB exposure rather than temperature, making this one of the first studies to document a negative effect of organochlorine contaminant exposure in elasmobranchs in a field setting. Our results suggest that developing embryos are sensitive to maternally offloaded in a sex-specific manner. Differential response to contaminant exposure indicates that life history characteristics should be considered as part of future studies. Currently, contaminant exposure and impacts are not factored into species management plans. Our results, measured in a small elasmobranch species, may have

36 reproductive implications for more difficult to study elasmobranchs, which are typically larger, more highly contaminated and possibly just as sensitive to contaminant exposure as the model used in this study. The unique and varied matrotrophic strategies of elasmobranchs makes them an interesting taxon to continue exploration into the relationship between maternal provisioning, offspring conflict, and the impacts that early exposure has on lifetime fitness.

Chapter 3: Developmental Embryonic Steroid Contributions to Histotroph in Round Stingrays Steroid hormones in elasmobranch (sharks, skates and rays) reproductive biology have generally been examined from the maternal perspective. Despite a wide range of elasmobranch matrotrophic strategies, and thus potentially diverse pathways for maternal steroid hormones to reach embryos, little attention has been given to uterine hormones. The objective of this study was to characterize steroid hormone profiles simultaneously in mother’s plasma and uterine fluid (histotroph) over the course of pregnancy with respect to developmental landmarks in embryos. Round stingrays were sampled every month of their gestational cycle from post-ovulation to near parturition and pairs of samples of plasma and histotroph were analyzed for a suite of steroid hormones using LC-ESI/MRM. Hormone concentrations were compared within and between maternal and uterine compartments using two parameters of embryo development. Histotroph had consistently higher detection rates and concentrations of hormones than plasma, especially during early pregnancy (embryos < 10g). Furthermore, histotroph peaks in testosterone preceded plasma, suggesting that hormones were locally produced. Finally, the peak in histotroph hormone concentrations coincided with embryo sexual differentiation based on the presence of visible claspers (male copulatory organs). Results suggest that, like mammalian pregnancy, embryonic steroids contribute to the developmental environment.

3.1 Introduction Elasmobranch (i.e. sharks, skates and rays) reproductive endocrinology has focused on measuring maternal plasma hormone concentrations to determine the timing of landmark events such as mating, ovulation, and parturition. While this information is necessary for proper fisheries management, it is too coarse to identify steroid hormone communication between mothers and embryos that may be occurring more locally. Many elasmobranch species are viviparous, exhibiting a wide variety of matrotrophic provisioning strategies ranging from pseudoplacental viviparity to oophagy to histotrophy (Hamlett et al., 1993; Wourms et al., 1988). It follows that the diversity of provisioning pathways would allow maternal hormones different routes by which to reach embryos. In mammalian pregnancies, maternal recognition of the pregnancy is important for the prevention of embryo expulsion and the creation of an appropriate environment for developing

38 embryos. Steroid hormones play key roles in initiating these changes by quieting uterine contractions and altering uterine secretions. In swine and lagomorphs, blastocysts are steroidogenic as early as 14 and 17 days post-fertilization, respectively (Bazer et al., 1979; Wilson et al., 1980). These embryo-produced signals alter the secretory functions of the uterus related to glucose and protein transport as well as triglyceride synthesis (Forde et al., 2009). Mammalian embryonic sex steroids produced in utero can also have long lasting, downstream developmental effects even into adulthood (reviewed in Zambrano et al., 2014). In sharks and rays, where maternal investment in pregnancy can be high as it is in mammals, both maternally-derived and embryo-derived steroid hormones are expected to be important for reproductive success as well. Round Stingrays Urobatis halleri have several characteristics that make them ideal as an elasmobranch model. They are abundant, easy to sample and have a well-described reproductive cycle (Babel, 1967; Hoisington and Lowe, 2005; Mull et al., 2010). In addition, Round Stingrays employ histotrophic matrotrophy to provide supplemental nutrition to embryos, whereby the uterus produces protein and lipid-rich secretions that nourish embryos throughout development, particularly after they have exhausted their yolk sacs (Babel, 1967; Spieler et al., 2013). From the beginning of pregnancy, the uterus is densely lined with highly vascularized uterine villi (trophonemata) that transfer nutrients from the maternal compartment to the embryonic compartment through uterine secretions (i.e. histotroph). The close physical connection between these compartments may also have the potential to export steroid hormones from the histotroph into maternal circulation, as occurs in humans (Murphy et al., 2006). Progesterone has been detected in histotroph from the Marbled Electric Ray Torpedo marmorata (Fasano et al., 1992), but the stage of pregnancy was not reported. Thus, there is a knowledge gap about histotroph steroids and the temporal pattern of change over development. The objective of this study was to investigate how steroid hormones change over the course of pregnancy within maternal (plasma) and histotroph (uterine fluid) compartments. Specifically, I was interested in the alignment and timing of histotroph concentrations with developmental events (i.e. sexual differentiation, absorption of yolk sacs), and the association, if any, between maternal plasma and histotroph concentrations. Based on established patterns in mammalian pregnancy, I hypothesized that early in development, histotroph hormones would reflect maternal plasma, but that sexual differentiation would be associated with embryonic steroidogenesis.

3.2 Methods 3.2.1 Study populations. Female stingrays were sampled from two locations in southern California: mainland California at the Seal Beach National Wildlife Refuge (SBNWR, 33.731N, 118.064W) and Catalina Harbor, Catalina Island (33.434N, 118.503W). Adult males from both study areas were also collected but were infrequent in the bays with the pregnant females. Thus, males were collected from along the open coast at Seal Beach (33.739N, 118.113W) at the mainland site, and at Two Harbors (33.443N, 118.499W) from the island site. The sites differ in anthropogenic contamination, with established effects on embryonic development for the same pregnancies as are reported here (Chapters 2, 5, 6). Specifically, polychlorinated biphenyls comprised a majority (≥ 75%) of accumulated organic contaminants. Other potent environmental disrupting chemicals, such as dichlorodiphenyltrichloroethane and pharmaceuticals, were low (Chapter 2) or assumed to be none detectable based on previous work (Lyons unpublished data), respectively, when the nearshore ecology of the species is considered with regards to the offshore release of treated wastewater (~8km from the coast, > 200m deep). We found no indication that contaminant exposure had an effect on hormone concentrations for either plasma or histotroph. Rather, asynchrony in reproductive timing was more important (see below). Temperature also differed at the two sites where females were sampled, influencing reproductive timing by promoting earlier ovulation in females from the contaminated site. In spite of warmer waters and earlier commencement of pregnancy, embryos from the contaminated site grew more slowly, and were smaller at every developmental stage, than embryos from the cleaner, island site. Differences in embryo mass between sites were sex-specific, with contaminant-exposed embryonic males being significantly lighter than their reference (island) counterparts and embryonic females showing no differences. Due to embryonic differences in growth, embryos were aligned using a developmental marker (clasper appearance), which is described below and was the optimal method, with the information available, to account for site differences in growth. 3.2.2 Sampling. Females were sampled approximately monthly across their pregnancy (Jirik and Lowe, 2012), to include reproductive landmarks such as post-ovulation, early development with embryos still yolk dependent, mid-development embryos exhausting their yolk sacs, and late development

40 with embryos completely dependent on histotroph. Pregnant females and adult males were captured via hook and line and subjected to one of two treatments (“baseline” and “stressed”) in order to also examine the effect contaminant exposure had on the acute stress response (Chapter 4). The short duration of the stressor (~15 mins) was not expected to alter blood or histotroph reproductive steroid hormone levels, and there was no evidence of loss of histotroph hormones as uterine flushing did not occur (Chapter 5); thus, stress was not considered a factor in these analyses. After capture, stingrays were euthanized with an overdose of tricaine methanesulfonate (MS-222) in accordance with animal care protocols (University of Calgary Animal Care Protocol #14-0016). Prior to tissue sampling, disk width (a measure of size) of all females was taken. Maternal blood samples were obtained via cardiac puncture, transferred to a heparinized tube, and stored on ice until components could be separated via centrifugation at the lab. Plasma was then stored at -80°C until analysis. After blood sampling, the skin over the ventral side of the body was carefully removed and internal organs (i.e. liver and digestive system) were excised to expose both uteri. For all females, a small slit was made in the left uterine wall and a 10 mL syringe with an 18 gauge needled was used to aspirate as much histotroph as possible. Histotroph was also collected from the right uterus in four females. After transfer to a 15 mL tube, histotroph was then immediately frozen on dry ice and subsequently stored at -80°C until analysis. Embryos were then removed from both left and right uteri, individually wrapped in aluminum foil, and frozen for later dissection, where the sex (determined by the presence or absence of visible claspers, the male copulatory organ), disk width, total body mass and weight of yolk sac (if present) were recorded. 3.2.3 Hormone quantitation. 3.2.3.1 Chemicals and reagents. Cortisol, cortisone, corticosterone, 11-deoxycortisol, 11-dehydrocorticosterone (11-DHC), testosterone, 17α-OH-progesterone, androstenedione, progesterone, 11-ketotestosterone, estrone, estradiol and estriol were purchased from Steraloids Inc (Newport, RI). Deuterium labeled internal standards: cortisol-d4, corticosterone-d8, testosterone-d2, 17α-OH-progesterone-d8 and progesterone-d9, Estrone-2,4,16,16-d4, 17β-estradiol-2,4,16,16-d4 (estradiol-d4), 16α-Hydroxy- 17β-estradiol-2,4-d2 (estriol-d2) were obtained from CDN Isotopes Inc (Pointe-Claire, Quebec, Canada). HPLC grade methanol, Optima grade acetonitrile and Optima grade water were purchased from Fisher Scientific (Edmonton, AB, Canada). While 1α-OH-corticosterone is the

41 predominant corticosteroid used by elasmobranchs (Anderson, 2012), the unavailability of reference standards precludes it from LCMS quantitation, although a corticosterone enzyme immunoassay is available (Evans et al., 2010). To maintain method consistency, I did not include 1-α-OH-corticosterone as part of these analyses. 3.2.3.2 Sample preparation procedure. A 50 µL sample of maternal plasma or histotroph was transferred into a 0.5 mL micro- centrifuge tube, followed by 50 µL of protein precipitation solution (ZnSO4•7H2O at 9 mg/mL, in methanol and containing deuterated internal standards [IS]). After 20 mins of cold incubation (10˚C), the mixture was vortexed for 15 s, centrifuged at 14,000 rpm for 15 mins, and the 50% water/50% methanol supernatant (75 µL) was submitted for LC-MS analysis. 3.2.3.3 LC-ESI/MRM analysis. Pooled samples of maternal plasma and histotroph were used to establish calibration ranges and as quality control (QC) in the sample matrix. All samples were analyzed using an Agilent 1200 binary liquid chromatography (LC) system coupled to an AB SCIEX QTRAPÒ 5500 tandem mass spectrometer equipped with an electrospray ionization (ESI) source. LC separation was performed on an Agilent ZORBAX Eclipse plus C18 column (100 x 2.1 mm, 1.8 µm particle size) at 40˚C for Positive ESI or Agilent Poroshell 120 C18 column (50 x 2.1 mm, 2.7

µm particle size) at 35˚C for Negative ESI. The mobile phase A was ACN/H2O (5/95, v/v, 2mM NaF) and the mobile phase B was 100% ACN (2mM NaF). The 12 min gradient for Positive ESI mode was 15-70% B (0-6 min), 70-100% B (6-7 min), 100% B (7-8.5 min), 100-15% B (8.5-9 min), held at 15% B for 3 min. The 9 min gradient for Negative ESI mode (estrogens) was 40- 80% B (0-4 min), 80-100% B (4-4.5 min), 100% B (4.5-5.5 min), 100-40% B (5.5-6 min), held at 15% B for 3 min. The flow rate was 0.4 mL/min and the injection volume was 12 µL. Mass resolutions in Q1 and Q3 were set to unit resolution. Each analyte was monitored by two transitions (a quantifier and a qualifier; Table 3.1). 3.2.3.4 Quantitation method. Steroids were quantified as area ratio relative to the appropriate deuterated internal standard. Calibration curves used analyte/IS peak area ratios (y-axis) vs. analyte concentration (x- axis) and a linear fit with 1/x weighting. Fit (i.e. r2) for all calibration curves were always higher than 0.995, and typically higher than 0.998. The lowest concentration of a steroid that gives < 20% coefficient of variation (CV) and < ± 30% error was determined as its lower limit of

42 quantification (Table 3.13). The system of quantification allows us to account for potential matrix influences within samples. Three quality control solutions were run in duplicate. At the lowest QC tested (0.25 or 0.5 ng/mL), duplicate CV for individual steroids (mean CV across steroids = 5.72%) ranged from 0.1% (deoxycortisol) through 15.1% (estradiol). Quantified hormone concentrations (ng/mL) were converted to nmol/L prior to analyses. 3.2.4 Data analysis. R statistical package (V 2.13.0) (Team, 2011) was used for all analyses and α was set to 0.05. Mean ± the variance is represented as standard deviation (SD), and % CV (SD/mean*100) is reported for quantitation duplicates. Given the previously established reproductive asynchrony between populations (Chapter 2), litters were grouped based on either developmental stage or sexual differentiation. Both measures align pregnancies at the two locations, so that site was not an analytic variable. For developmental stage, litters were assigned to one of the following groups: post-ovulation (no visible embryos), early-term (embryos yolk dependent), mid-term (embryos exhausting yolks), and late-term (embryos completely histotroph dependent)(Table 3.2). For sexual differentiation, litters were divided into pre- and post- groups based on clasper appearance and sites were aligned by clasper days (Chapter 2). Clasper days represented our best way of aligning litters, although we acknowledge that we do not know the effect that temperature or degree of maternal investment plays on the first appearance of claspers. In this way, I investigated hormone changes with respect to yolk utilization as well as sexual differentiation. Hormonal differences were explored using the appropriate multiple comparisons tests and linear regressions, respectively. Maternal and embryonic compartments were examined separately before pairwise comparisons of plasma and histotroph were conducted using Pearson’s correlations and paired t-tests. No attempt was made to quantify yolk steroids. However, to determine the possibility of the yolk sac possessing steroidogenic activity, thus representing a potential source of hormone production in the histotroph, hormone concentrations were compared relative to yolk sac mass and embryo mass. Yolk sac size was expressed as a percentage of total embryo mass (i.e. [yolk sac weight]/[yolk sac weight + embryo weight] *100). If yolks are the source of histotroph hormones, where hormones are released into the uterine environment directly from the yolk rather

3 http://www.absciex.com/Documents/Downloads/Literature/mass-spectrometry-cms_059150.pdf

43 than routing through the embryo, I predicted that histotroph hormone concentrations would reflect the same patterns as the decrease in yolk sac mass following reabsorption.

Table 3.1. Mass transitions and limits of quantitation for all steroids. The quantifier (left) and qualifier (right) mass transitions of targeted steroids and their lower limit of quantification (LLOQ) are given. The total number of deuterium substitutions in the internal standard is shown in parentheses. Analyte Transitions Transitions LLOQ (ng/mL) (non-deuterated) (deuterated) Cortisol (d4) 363/121 363/327 367/121 367/331 0.1 Corticosterone (d8) 347/329 347/121 355/337 355/125 0.05 Cortisone 361/163 361/121 - - 0.05 11-dehydrocorticosterone 345/121 354/301 - - 0.05 Testosterone (d2) 289/91 289/109 291/99 291/111 0.05 17α-OH-progesterone (d8) 331/97 331/109 339/100 339/109 0.05 Progesterone (d9) 315/97 315/109 324/100 324/113 0.05 11-deoxycortisol 347/97 347/109 - - 0.05 Androstenedione 287/97 287/109 - - 0.05 11-ketotestosterone 303/121 303/259 - - 0.1 Estriol (d2) 287/171 287/145 289/173 289/147 0.1 Estradiol (d4) 271/145 271/183 275/147 275/187 0.05 Estrone (d2) 269/145 269/143 273/147 273/145 0.05

Besides changes with development, other factors such as litter sex composition or litter number could influence histotroph concentrations. Unfortunately, embryo position (i.e. right or left uterus) was not recorded, save for the four samples where histotroph was collected from both sides, and the anatomic expectation (Spieler et al., 2013) is that uterine fluid cannot be exchanged between uteri. Therefore, tests of effect of sex on histotroph steroid concentration used only the subset of litters that were single sex. I hypothesized that male-only litters would have greater testosterone and lower estradiol concentrations than female-only litters based on mammalian studies.

3.3 Results Our pregnant female sample set (n = 55) was comprised of post-ovulatory (n = 6), early- (n = 11), mid- (n = 14) and late-term (n = 24) females. A total of 14 adult male stingrays were captured in July at the contaminant-exposed site, and 7 males were captured in August at the

44 island site. As some early-term females aborted their litters before histotroph was sampled (n = 6), histotroph-plasma pairs were not available for these samples. After correcting for previously established reproductive asynchrony (Chapter 2), no additional effects of site were identified. Therefore, samples were assessed with sites combined for analyses. Pooled samples of maternal plasma and histotroph were run in triplicate for quality control assessent. Only cortisol and testosterone were quantifiable in pooled maternal plasma (cortisol CV = 7.0%; testosterone CV = 6.1%). In pooled histotroph, cortisol was below LOQ, but testosterone (CV = 4.0%), 17-OH progesterone (CV = 9.7), progesterone (CV = 4.6%), estriol (CV = 3.7%) and estradiol (CV = 2.5%) were reliably quantifiable. 3.3.1 Males. For adult males, testosterone was quantified in all samples (contaminant-exposed/July 5.86 ± 3.65 nmol/L, reference/August 7.75 ± 4.62 nmol/L) and did not differ between sites (p = 0.4). At the contaminant-exposed site, 4/14 males (28.6%) had detectable cortisol in plasma (6.04 ± 11.3 nmol/L when above LOQ) whereas no (0/7, 0%) reference males had detectable cortisol. I found a weak effect of site on the probability of detecting cortisol (χ-square =2.47, df = 1, p = 0.058). At the contaminant-exposed site, the only other steroid detected was progesterone in two males (0.31 – 8.14 nmol/L). At the reference site, one male had both progesterone and 11- dehydrocorticosterone detections (0.6 and 0.15 nmol/L) and a second male only estradiol (0.47 nmol/L) in addition to testosterone. 3.3.2 Maternal Plasma. As predicted by the pooled maternal plasma steroid quantitation, testosterone was consistently detected in female plasma (54/55, 98%). Estradiol (17/55, 31%), progesterone (7/55, 13%), and 11-dehydrocorticosterone (4/55, 7%) were detected less frequently. Unlike the pooled plasma, cortisol was rarely quantifiable (5/55, 9%). Additional hormones quantified in this study were not above LOQ in any plasma samples. Maternal plasma testosterone was the only hormone to show strong development-related changes in concentration (KW, W = 21.5, p < 0.0001). Testosterone concentrations increased from post-ovulation to peak mid-term before subsequently decreasing in late-term females (Figure 3.1). Not surprisingly, adult males had significantly higher testosterone concentrations than pregnant females at both the contaminant-exposed

(Welch’s t15 = 2.95, p = 0.005) and reference site (Welch’s t6 = 2.95, p = 0.013; Figure 3.1). When females were compared to males at the same collection time period (i.e. July for

45 contaminant-exposed and August for reference), contaminant-exposed males no longer had higher levels than females, due to the coinciding peak in female testosterone during that time (p = 0.09). Estradiol was most frequently detected in post-ovulatory females (3/6) followed by late (9/24) and early (4/11) term females, with only one mid-term detection. When quantified, estradiol marginally increased from early (0.24 ± 0.049 nmol/L) to late-term pregnancy (0.44 ± 0.32 nmol/L; p = 0.05). Progesterone had single detections in both the post-ovulatory and early- term time periods, with the majority of detections occurring during the late term (5/24). Progesterone was highest in a post-ovulatory female (1.14 nmol/L), but similar between an early- term (0.19 nmol/L) and late-term females (0.18 ± 0.028 nmol/L). Cortisol was only detected in post-ovulatory females (4/6) and one early term female, whereas 11-dehydrocorticosterone was only detected in late term females (4/24). The lack of detections of cortisol during the same sampling time as males, suggests that different sex-specific factors influence the presence of this hormone.

Figure 3.1. Testosterone concentrations in maternal plasma by pregnancy stage and site. Females showed developmental stage-related changes in plasma testosterone (A), with concentrations peaking during mid-pregnancy as yolk is almost exhausted. Over all samples, females (light grey) had significantly lower concentrations than males (dark grey) at both the contaminant-exposed and reference site (B). Numbers in parentheses represent samples with detections/total samples and dashed line denotes the lower limit of quantitation for testosterone.

3.3.3 Histotroph. Confirming the histotroph findings in the Marbled Electric Ray (Fasano et al., 1992), progesterone was quantifiable in every sample (49/49, 100%). Save for testosterone, which was comparable to plasma (48/49, 98%), hormones generally had greater detection rates in histotroph than plasma. 17α-OH-progesterone (OHP; 41/49, 84%), estriol (28/49, 57%), and estradiol (25/49, 51%) were quantifiable throughout pregnancy. On the other hand, androstenedione (13/49, 27%) and estrone (6/49, 12%) were detected in only early and mid-term samples, with the former measured in all nine early-term litters. With respect to corticosteroids, 11-deoxycortisol, a hormone upstream of cortisol in the synthesis pathway, was detected throughout pregnancy (15/49, 31%), while “endpoint” corticosteroids (i.e. corticosterone and cortisol) were detected in mid and late term litters (7/49, 14% and 2/49, 4%, respectively; Table 3.2). Only four hormones showed concentration changes with development (progesterone, testosterone, OHP, and estradiol). With respect to developmental stage, patterns of increase and decrease were most similar between progesterone and testosterone (Figure 3.2). Concentrations increased sharply from post-ovulation to peak in early-term litters, followed by weak decreases in mid-term litters, with lowest levels in late-term litters (KW, W = 36.5, p < 0.0001 and W = 27.1, p < 0.0001, respectively). OHP also peaked in early-term litters but concentrations decreased more gradually as mid-term concentrations were not significantly different than late-term (KW, W = 13.6, p = 0.001; Figure 3.2). Despite estradiol being detected less consistently, concentrations showed similar patterns as the other three hormones with highest concentrations in early-term and lowest during late-term (Welch’s t7 = 2.5, p = 0.02; Figure 3.2). Therefore, embryonic changes occurring during early-term development alter histotroph concentrations. Based on clasper days, I also found changes in hormone concentrations over development (Figure 3.3). Males and females were distinguishable at a small size (approximately 3.2% and 3.5% of final mass), with sexual differentiation occurring during the early-term stage. Hormones tended to peak just before or at the first appearance of claspers (i.e. clasper day 0), and subsequently fell during the post-differentiation phase. Progesterone showed the strongest increase (p = 0.047, r2 = 0.41) and decrease (p < 0.0001, r2 = 0.37) in hormone concentration within each phase. Changes in OHP mirrored that of progesterone; however, increases and decreases were weaker, especially during post-differentiation. With respect to sex-related hormones, estradiol concentrations peaked just prior to the appearance of claspers, while

48 testosterone was highest at clasper day 0. During the post-differentiation phase, testosterone rapidly fell in concentration (p = 0.0002, r2 = 0.32) like progesterone, while estradiol only showed weak decreases (p = 0.06). The peak in hormone concentrations around the time of clasper appearance (or lack thereof in females), implicates embryo sexual differentiation as the developmental landmark responsible for altering the histotroph steroid hormone environment during early development. The potential for embryo yolk sacs to serve as a source of steroid hormones to the embryo compartment was examined by comparing concentrations to relative yolk mass (sites combined). During early development, yolk sacs were rapidly exhausted and were completely absent after embryos had reached ~20% of their final mass. As yolk sacs were depleted, concentrations of progesterone and testosterone increased in histotroph during early development (Figure 3.4).

Table 3.2. Histotroph Steroids by stage of pregnancy. Mean ± SD concentrations (nmol/L) of steroids in histotroph, with litters divided into four developmental stages (sites combined): post-ovulatory (eggs, no embryos), early-term (embryos fully dependent on yolks), mid-term (embryos exhausting yolks), and late-term (embryos fully histotroph-dependent). Number of pregnant females and associated histotroph collected (in parentheses) is given for each group. Cortisone and 11-ketotestosterone were not detected in any samples. Post- Early-term Mid-term Late-term ovulatory Cortisol - - - 0.50-0.69 11-deoxycortisol - 0.20 ± 0.047 0.27 ± 0.16 0.22 ± 0.06 11- - - - 0.36 dehydrocorticoste rone Corticosterone - - 0.57 ± 0.13 2.75 ± 2.22 Testosterone 3.54 ± 1.52 27.7 ± 22.5 16.9 ± 12.5 3.30 ± 1.87 17α-OH- 0.83 ± 0.38 2.70 ± 0.86 1.74 ± 1.06 1.09 ± 0.90 progesterone Androstenedione - 0.68 ± 0.43 0.42 ± 0.36 0.17 Progesterone 3.69 ± 2.40 85.2 ± 50.9 35.4 ± 22.4 4.82 ± 3.81 Estriol 1.51 ± 0.26 2.83 ± 0.36 2.49 ± 0.23 1.53 ± 0.42 Estradiol - 1.02 ± 0.72 0.42 ± 0.24 0.65 ± 0.38 Estrone - 2.56 ± 2.26 1.93 ± 1.16 0.55 ± 0.26 # Litters 6(3) 11(9) 14(13) 24(24) (histotroph) Developmental Stages

Figure 3.2. Histotroph progesterone, 17α-OH-progesterone, testosterone and estrogen changes with pregnancy stage. Histotroph was compared among litters based on developmental stage, with sample detections given in parentheses, for progesterone (A), testosterone (B), 17α-OH-progesterone (“OHP”, C) and estradiol (D). Different letters denote significant differences among groups (α < 0.05). The absence of a letter indicates groups with too few samples for statistical comparison. Dashed horizontal line indicates lower limit of quantitation for each hormone. Note that concentrations for progesterone and testosterone are on log scale.

200 5 Pre Post Reference Pre Post Reference Exposed A Exposed C 4 150

3 100 2 OHP (nmol/L) OHP

Progesterone (nmol/L) 50 1

0 0 -50 Eggs 0 50 100 -50 Eggs 0 50 100 only Clasper Day only Clasper Day

B D 80 Reference 8 Pre Post Exposed Pre Post Reference Exposed

60 6

40 4 Estradiol (nmol/L)

Testosterone (nmol/L) 20 2

0 0 -50 Eggs 0 50 100 -50 Eggs 0 50 only Clasper Day only Clasper Day

Figure 3.3. Histotroph steroid changes at the time of external sexual differentiation. To compare how histotroph hormone concentrations changed with respect to sexual differentiation, I divided litters into pre- (light grey) and post- (dark grey) groups and aligned between sites using clasper days as a developmental marker. Litters were pooled between contaminant-exposed (triangles) and reference sites (circles) for analyses and significant relationships are depicted with solid lines and marginally insignificant ones with dashed lines. The first appearance of claspers on male pups (0 clasper day) is denoted by a vertical dashed line, while dashed horizontal line indicates lower limit of quantitation for each hormone. Note that all values are on log scale and estradiol in post-ovulatory adult females was below detection limits.

Figure 3.4. Impact of yolk on histotroph progesterone and testosterone concentrations. The absorbance of yolk sacs over early pregnancy, depicted as the contribution of yolk mass to total embryo plus yolk mass, is compared to concentration changes of progesterone (triangle) and testosterone (circle) and embryo mass (squares). During early development when yolk mass contributes the most (> 50%) to total mass there is a spike in hormone concentrations. However, hormone concentrations rapidly decline after sexual differentiation (vertical dashed line) despite the continued utilization of yolk sacs.

3.3.4 Maternal vs. embryo. As testosterone was consistently detected in plasma and histotroph, I used it to test our hypothesis that hormones would be maternally derived in early development and embryo derived later in development. Pairs of plasma and histotroph showed stage-specific correlations (Figure 3.5). No correlations existed for early-term litters (p = 0.1), but samples were correlated in mid- term litters (p = 0.019, r = 0.58). By late-development, plasma-histotroph pairs had become only weakly correlated (p = 0.043, r = 0.36). No correlations existed between plasma and histotroph progesterone and estradiol, regardless of stage. Contrary to predictions, histotroph testosterone was significantly higher than its corresponding plasma pair for all stages for early (t7 = 3.0, p =

0.019), mid (t12 = 3.9, p = 0.002), and late-term (t22 = 2.4, p 0.035) litters (Figure 3.5). Dividing litters by sexual differentiation lead to similar findings of histotroph having significantly higher concentrations than paired plasma (pre: t7 = 2.6, p = 0.034; post: t37 = 3.3, p = 0.002). Testosterone peaks in each respective tissue also differed (Figure 3.5). Histotroph preceded plasma and peaked during early-term development, whereas plasma testosterone did not peak until the mid-term. Therefore, contrary to our hypothesis, histotroph influences plasma concentrations both earlier than predicted and more strongly than plasma influences histotroph. To examine the influence of sex, I compared litters that had only males or females (n = 5 and 6, respectively). Previous work noted site-related differences between male stingray embryos (Chapter 2). Unfortunately, low sample size prevented us from determining if there were significant differences between sites with respect to sex hormones. As most of the samples were from the contaminant-exposed site (7/11), I was able to compare concentrations based on mean litter mass. Testosterone was comparable in early term litters regardless of sex (Figure 3.6) as was progesterone. Testosterone decreased more steeply in female-only litters after differentiation. Conversely, estradiol and estriol were detected in female-only litters much earlier than in male-only litters (Figure 3.6). When estrogens were detected in male-only litters, concentrations were generally comparable to that of females. The earlier presence of estradiol in female-only litters suggests possible sex-related influences on histotroph composition.

Figure 3.5. Associations between maternal plasma and histotroph testosterone. (A) Testosterone in plasma and histotroph pairs were not correlated in early-term litters (p = 0.1, open circles), but were correlated (p < 0.02) during the mid-term (grey triangles) when plasma testosterone peaked. By late-term development (black squares), weak correlations existed between plasma and histotroph pairs (p = 0.04). (B) In all developmental stages, histotroph had significantly higher concentrations than plasma pairs. Histotroph concentrations peaked prior to plasma. Cubic spline curves were fitted to the data to depict concentration changes with development. Among possible sources of error are that testosterone concentrations could not be corrected for the number of male embryos in the left uterus, where the histotroph was collected.

Figure 3.6. Impact of sex-biased litters on histotroph steroids. To investigate potential sex-related effects on histotroph steroid composition, I compared concentrations of male-only (n = 5, triangles) to female-only (n = 6, circles) mid-late development litters. (A) Testosterone (solid lines) and progesterone (dashed line) were similar between the sexes. (B) Estradiol (solid lines) and estriol (dashed lines) were both detected earlier in pregnancy in female-only litters compared to males, suggesting a sex-related effect on histotroph. Vertical dashed line indicates the size at which litters could be sexually differentiated, and horizontal lines denote the lower limit of quantitation.

3.4 Discussion 3.4.1 Adult plasma sex steroids. Not unexpectedly, males had significantly higher concentrations of testosterone than females when sampled in the summer, and these differences would likely be much larger when males are at the height of testicular activity in the fall (Mull et al., 2008). Other studies have noted the presence of testosterone in elasmobranch females (Manire et al., 1995), particularly peaking around mating and ovulation. These latter reproductive milestones were not sampled in the present study. However, we found female testosterone to peak much later than previous elasmobranch studies. The increase in testosterone later in pregnancy is likely due to uterine influences on plasma (see below), and differences in matrotrophic strategy between Round Stingrays and placental sharks that may also influence testosterone peaks in maternal plasma. I also found no difference in testosterone concentration between contaminant-exposed and reference males, despite previous work implicating PCBs as having negative effects on male reproduction (Frantz, 2014), and established differences between the sites for PCB exposure (Chapter 2). Unfortunately, there was covariance of sampling date and site for the males, which makes comparisons challenging. Further studies exploring male reproductive effects should be conducted during the fall when testosterone concentrations peak (Mull et al., 2008), allowing any site differences to be most visible at this time. The dearth of hormone detections in female plasma prevented rigorous testing for the effects of PCBs on female steroidogenesis. However, by comparing females based on embryo development, we were able to account for differences in environmental temperature, which did have an effect on reproductive timing. In both populations, I was able to quantify a variety of corticosteroids, including one that has not been reported in elasmobranchs (11-deoxycortisol), despite it being the main corticosteroid of the sea lamprey Petromyzon marinus (Close et al., 2010), a jawless fish. Unfortunately, without new method development, I was unable to measure the main corticosteroid utilized by elasmobranchs (1α-hydroxycorticosterone) via LCMS. However, the effect of stress on 1α-hydroxycorticosterone was quantified by immunoassay and compared among adults and in Chapter 4. While I quantified corticosteroids in both sexes (this study), males and females did not overlap temporally in their detections of these hormones. As the mating season concludes in May, endocrine differences between males by site may be indicative of differential male-male competition for resources as the population at Catalina is purportedly

57 smaller, and likely less dense, than the mainland (Plank et al., 2010). Temperature was not expected to significantly influence male-male comparisons as water temperature at the two sites where males were sampled were similar (Chapter 2). For females, studies have noted increases in corticosteroids during ovulation in other elasmobranchs (Manire et al., 2007), and likely accounts for the presence of cortisol in island female stingrays post-ovulation. Our inability to sample ovulating females at the mainland prevented us from making site comparisons. 3.4.2 Mothers: hormone sink or source? The low concentrations and detection rate of hormones in plasma was unexpected, considering that, in other elasmobranchs, concentrations of sex steroids tend to peak around the time of ovulation and are detectable throughout pregnancy (Manire et al., 1995; Mull et al., 2010; Tricas et al., 2000). The reduced number of detections could be related to the sensitivity of our method (0.05 - 0.1 ng/mL) compared to other studies (3 pg/mL; Manire et al., 2007) as well as the solvents employed since steroid structure influences its solubility. Histotroph concentrations were higher, and more readily detected. One explanation could be that steroidogenesis in adult females substantially declines after ovulation compared to other species, resulting from inherent differences in their reproductive biology. Round Stingrays have a fast gestation period (three to four months) compared to five to 18 months in some species. Assuming that sexual differentiation occurs at the same relative size, the slower development in other species would result in longer time periods between the shift from maternal to embryo hormone production. Histotroph may represent a hormone source during pregnancy while maternal plasma represents a sink. Therefore, the low concentration and detection rate of plasma hormones could be a result of maternal metabolism of uterine-derived hormones rather than low ovarian production. I had predicted hormone transfer between maternal (plasma) and embryonic (histotroph) compartments would occur relatively quickly due to the close proximity of these two tissues aided by the thin villi of the trophonemata. However, plasma and histotroph testosterone concentrations were only strongly correlated when plasma testosterone peaked. 3.4.3 Yolk as a source. Hormones present in histotroph could result from the release of those laid down in yolk tissue by mothers during vitellogenesis, which are then liberated as embryos consume yolks. Particularly in egg-laying animals, females supplement developing ova with chemical signals

58 that maternally influence embryonic development in the absence of a physical maternal-embryo connection (Adkins-Regan et al., 1995; Conley et al., 1997; Schwabl, 1996). In elasmobranchs, some steroid (Manire et al., 2004) and thyroid hormones (triiodothyronine and thyroxine)(McComb et al., 2005) have been previously measured in the yolks of Bonnethead Sharks (Sphyrna tiburo) from pre-ovulation through early embryonic development. McComb et al. (2005) suggested that increases in yolk thyroid hormones could be due to conversion activities in the yolk as well as possible embryonic production. In birds and reptiles, hormone concentrations vary among different layers of yolk (Bowden et al., 2001; Moore and Johnston, 2008), which would alter their temporal availability. In the present study, the increase in histotroph progesterone, testosterone and OHP were likely not the result of hormone release into histotroph from yolks during reabsorption. Firstly, concentrations measured in histotroph with post-ovulated eggs had significantly lower concentrations of progesterone, testosterone, OHP and estradiol than histotroph containing visible embryos. In addition, hormones present in yolk should be first taken up by embryos before being released into the uterine environment (Conley et al., 1997), which would limit the rate of release of yolk hormones into histotroph. Simple diffusion of hormones from the yolk to histotroph to explain the rapid rise in concentration during early development also does not seem plausible. As lipid content of Round Stingray yolk is substantially greater than that found in aqueous histotroph (Lyons, unpublished data), I would expect slow direct transfer rates of lipophilic hormones in yolk to histotroph. Therefore, while yolks likely contribute to some portion of histotroph hormones, they are not the main driver behind the large increase and sudden decrease in hormone concentrations during early development. 3.4.4 Embryos as a source. The lack of strong correlations and significantly higher concentrations of hormones in histotroph compared to plasma suggests that hormonal changes are driven by embryo development and maternal influence, if any, is relatively short-lived. Other studies have documented the steroidogenic ability of early stage embryos (Stone et al., 1986; Wilson et al., 1980), which leads to steroids concentrating to greater levels in uterine fluid than in plasma (Rivarola et al., 1968). Swine blastocysts (14-16 days post-fertilization) were found to be steroidogenic and utilize progesterone originating from their mothers to produce androgens and estrogens (Bazer et al., 1979). Many of these observations occurred during a short time frame

59 with high steroidogenic outputs on the part of the embryo, and a similar phenomenon could be occurring in stingray development although the time frame is protracted. While I do not have enough data to ignore potential yolk influences, our results suggest that embryo steroidogenesis, particularly during early development, plays the major role in altering histotroph hormone concentrations. Since egg yolks are known to contain cholesterol (Thompson and Speake, 2002), it is more likely that the yolk of Round Stingrays functions as a source of steroid precursor. I propose that embryos use yolk cholesterol for their own steroidogenic production; however, further testing is needed to verify this hypothesis. Progesterone is produced first, which would explain its high concentration, followed by downstream hormones, with testosterone likely being one of the main end-point hormones in early development. While, hormones originating from other tissues with steroidogenic activity (i.e. interrenal glands) were detected, particularly 11-deoxycortisol and corticosterone, their concentrations were much lower and detections less consistent, suggesting that gonadal over interrenal steroidogenesis is more of a priority in stingray embryo development. 3.4.5 Role of uterine hormones. The rapid, and significant, rise and fall of a particular set of hormones suggest that histotroph concentrations are driven primarily by embryos. It is compelling that the peak in these hormones occurs just prior to sex differentiation. In other vertebrate species, sexual differentiation is directed by the embryo rather than exogenously by the mother (Radder, 2007). It is not unlikely that the same regulation could be occurring in stingrays, although this is suggested with caution. In line with other taxa (Janzen et al., 1998) and elasmobranchs (Hoff, 2009; Manire et al., 2004), sexual differentiation occurs very early in Round Stingray development. The peak in estradiol and testosterone just prior to, and with, the appearance of claspers (male copulatory organs), respectively, suggest a role in sexual differentiation. Other intermediary hormones, such as OHP and androstenedione, may also promote this process (Glickman et al., 1987). While the presence of specific sex chromosomes is ambiguous in elasmobranchs (de Souza Valentim et al., 2013; Maddock and Schwartz, 1996), the results of this study suggest that testosterone, with its high detection rate and concentration, is a leading hormone candidate for the identification of sex-determining genes in elasmobranchs. Furthermore, the co-occurrence of estradiol during pre-

60 differentiation suggests testosterone may not be the only hormone responsible for sexual differentiation, and warrants further investigation. Besides serving as a precursor for sex steroid production, progesterone also likely plays a role in advancing the development of the uterus and quieting contractions to prevent embryo expulsion early in pregnancy. In mammals, embryonic-derived synthesis of progesterone has been associated with increase in pregnancy success in ruminants (Mann and Lamming, 1999) as well as inducing important changes in the uterus to provide nutrients to developing embryos (Spencer et al., 2008). In cattle, progesterone increases immediately post-conception and rises up to day 7, a critical window where increased progesterone concentrations are associated with increased embryo survival (Forde et al., 2009). The authors also documented many progesterone-induced gene changes in the endometrial lining, particularly for genes related to triglyceride synthesis, glucose secretion into histotroph, protein trafficking and transport. Since stingrays also produce histotroph, the high levels of progesterone measured in histotroph in early development may play a role in preparing the uterus for mid to late pregnancy when embryos have exhausted their yolk sacs and exclusively rely on uterine secretions (Hamlett et al., 1993). Conversely, while high progesterone is beneficial for early development, a fall in progesterone concentration might be required for expulsion of embryos at full-term. 3.4.6 Litter sex composition. I did not find any evidence that the sex of the litter influenced testosterone or estradiol concentrations; however, our sample size was small, and not focused around the interval when hormone concentrations peak. The only difference between single-sexed litters I observed was the detection of estradiol and estriol during early development in female-only litters but not in male-only litters, although these results should be interpreted with caution due to the limited number of single-sexed litters available. 3.4.7 Conclusions. Embryos alter their uterine steroid environment in Round Stingrays, with changes closely tied to sexual differentiation early in development. Future studies will need to record embryo sex and number independently for each uterus, and collect histotroph from both uteri to match to those embryos. With those data it should be possible to test the hypothesis that male embryos alter the sex steroid environment for females sharing the same uterus (Nagamani et al., 1979). I also recommend projects investigating sex-determining genes to consider screening for those

61 associated with androgen receptors (O’Shaughnessy et al., 2015), as this may yield promising results. Exploration in other elasmobranch species, such as skates, which develop independently of their siblings in their own egg cases, will also be interesting to examine litter sex composition effects. Thus, examination of the relationship between matrotrophic mode and embryo steroidogenesis, as well as characterizing steroid receptors in target tissues, will clarify the role, if any, for maternal-embryonic steroid signals in this ancient taxon.

Stress response impairment due to environmental contaminant exposure has been documented in a variety of organisms; however, little research has investigated these types of effects in elasmobranch fishes. Given the significantly higher PCB levels in Round Stingrays (Urobatis halleri) sampled from mainland California (contaminant exposed site) compared to those from a nearby offshore island (reference site), we tested the hypothesis that stingrays from the population with higher contaminant exposure would have a compromised stress response compared to stingrays from the reference population. Adult male and female stingrays were collected via hook and line from both locations and plasma was sampled either immediately or 15 min after a confinement stressor to measure parameters related to the primary stress response (i.e. corticosteroids) or secondary response (i.e. glucose). Our results showed potential impairment of the primary stress response in contaminant-exposed stingrays and a clear, significant effect of site on the secondary stress response, with both male and female stingrays from the contaminated site exhibiting no changes in their plasma glucose after stress but stingrays from the reference population demonstrating significant increases in plasma glucose post-stress. As a lower trophic level predator, the Round Stingray accumulates lower PCB loads than higher trophic elasmobranchs, suggesting that stress axis effects could be more severe in other elasmobranchs.

4.1 Introduction The ability to respond to stress, whether internally or externally derived, is critical for animal’s survival. When a stressor is recognized by the vertebrate central nervous system, both neuronal and endocrine responses are initiated (Axelrod and Reisine, 1984), and detection of a systemic stress response can be followed through the hierarchical changes that occur in blood parameters (Cockrem, 2013). The primary stress response is marked by a release of glucocorticoids into the blood (Bonga, 1997), which initiates a secondary response in key organs to enable the animal to cope with the stressor (Axelrod and Reisine, 1984; Mommsen et al., 1999). One important glucocorticoid function is to stimulate the liver to release glucose, an

63 important energy substrate used by most animals, into circulation (Mommsen et al., 1999). The release of glucose increases available energy that can be used by extrahepatic tissues (e.g. muscles, heart, brain) of an animal to cope with the stressor (e.g. fight or flight response). The liver is a crucial target of glucocorticoids as it is involved in many important metabolic functions and represents a storage cache of energy substrates such as glucose and glycogen that can be mobilized during stress. Measures of either the primary or secondary stress response are typified by a rise in specific plasma and tissue biomarkers (e.g. glucocorticoids, glucose) and subsequent fall after the stressor has been removed. However, the rate and magnitude of physiological parameter change in response to the stressor varies by species (Breuner et al., 2008; Cockrem, 2013) and among individuals within species. Disturbances to the stress response pathway can result from exposure to external agents, such as contaminants. Polychlorinated biphenyls (PCBs) are a class of legacy organic contaminants that were produced for industrial purposes, but whose production was banned after discovery that PCB exposure results in a multitude of negative physiological effects found across taxa (Hontela et al., 1992; Love et al., 2003; Sparling et al., 2010; Van den Berg et al., 2006). One well-documented effect of aquatic PCBs is the ability to disrupt the stress response cascade at multiple levels (i.e. brain, adrenals/interrenals, and/or liver). In aquatic ecosystems, PCB exposed Arctic Char (Salvelinus alpinus) have reduced expression of glucocorticoid receptors in the brain and lower levels of cortisol than fish not exposed to PCBs (Aluru et al., 2004). Similarly, in Rainbow Trout (Onchorhynchus mykiss) exposed to a PCB agonist, plasma cortisol levels and glucose concentrations were also depressed after exposure to an acute stressor (Aluru and Vijayan, 2006). Interference with the ability to mount an appropriate stress response will have adverse survival implications including a reduced ability to escape predation (Marentette et al., 2013) and to maintain homeostasis (Schreck, 2000). Therefore, all aquatic species prone to PCB accumulation might still be at risk. As predators, elasmobranchs (sharks, skates, and rays) tend to accumulate high levels of PCB contaminants (Fisk et al., 2002; Mull et al., 2013). However, the potential relationship between contaminant exposure and stress response impairment has not been closely examined due to the difficulty in quantifying the primary stress response in this taxon. Elasmobranchs, unlike other fishes, produce a unique corticosteroid (1α-OH-corticosterone) as their primary hormone to modulate metabolic and osmoregulatory functions (Anderson, 2012; Idler and

Truscott, 1966). However, synthesis of 1α-OH-corticosterone has proven difficult, resulting in the absence of a calibration standard. To circumvent this analytical problem, studies examining the stress response in elasmobranchs have focused on measuring variables related to the secondary stress response (i.e. changes in plasma glucose and ions; Hoffmayer et al., 2012; Manire et al., 2001; Marshall et al., 2012). These studies have primarily focused on post-capture survivorship (Skomal and Mandelman, 2012), so the potential for additional, subclinical effects of contaminants has not been explored. Studying the effect of PCB exposure in natural populations of elasmobranchs is particularly challenging due to their high vagility, making the availability of natural reference populations rare. In addition, many of the species of economic interest are relatively large and difficult to keep in captivity. The Round Stingray (Urobatis halleri), however, represents a practical model as it is abundant, easy to capture and has natural populations with different anthropogenic exposures. In particular, stingrays sampled from mainland California and an offshore island in southern California are known to differ in PCB accumulation (Lyons et al., 2014; Sawyna et al., 2017). Therefore, the objective of this study was to test the hypothesis that the acute stress response is attenuated in natural populations of stingrays exposed to environmental PCBs. Biomarkers of the primary and secondary stress response were measured at baseline, and after the application of an acute capture stressor in the contaminant-exposed and reference populations.

4.2 Methods 4.2.1 Sites. Stingrays were collected from two full seawater locations in southern California: mainland California, the contaminant-exposed site, and Santa Catalina Island, the reference site. Contaminant-exposed adult male stingrays were sampled from the shoreline at Seal Beach (33.739N, 118.113W) and pregnant females from within the nearby Seal Beach Wildlife Refuge (a protected estuary, 33.731N, 118.064W). Reference adult males were sampled from the leeward side of the island at Two Harbors (33.443N, 118.499W) and pregnant females were sampled from the windward side of the island in Catalina Harbor (33.434N, 118.503W; Figure 4.1). Stingrays from these two populations do not interbreed (Plank et al., 2010), thus stingrays at the reference island population are isolated from the contaminant-exposed mainland population.

Figure 4.1. Adult male and female sampling locations. Stingrays were sampled (black dots) from several areas within southern California. Seal Beach (Mainland California) represented our contaminated site and Santa Catalina Island, located ~35 km offshore, represented our reference site (males were sampled from the north side and females from the south side). Star indicates the location of the Palos Verdes US Environmental Protection Agency superfund site and relative depth is depicted in shades of grey, with darker colors corresponding to deeper depths. Arrows indicate direction of the California Countercurrent, reducing the environmental transport of contaminants from the mainland to the island.

The potential for genetic differentiation to alter the stress response of stingrays in the face of contaminant exposure is low. The genetic markers used to determine separation between sites are from non-coding regions of DNA, which do not provide any relevant measures of physiological differentiation. In addition, the period available for natural selection from when contaminant release peaked (1970) and the present study (2014) is relatively short. Considering that elasmobranch evolution occurs at a slower rate than mammalian evolution (Martin et al., 1992), based on mitochondrial DNA, genetic divergence is not expected to play a major role in physiological responses to contaminant exposure for each of our populations. A suite of organic contaminants was quantified in a subset of female stingrays (n = 8 from each site; Chapter 2), confirming previous findings that animals from the mainland site had significantly higher concentrations than those from the island (Lyons et al., 2014; Sawyna et al., 2017), with PCBs contributing to ≥75% of the total contaminant load. Round Stingrays prefer shallow, nearshore habitats (Babel, 1967), which spatially separates them from direct exposure to pharmaceuticals that could interfere with the stress response. In particular, treated wastewater in southern California is discharged into the offshore marine through benthic pipes located ~8km offshore and >200m deep. Indeed, common pharmaceuticals typically found in biota downstream of wastewater discharge were below detection limits in the hepatic tissue of Round Stingrays (Lyons unpublished data). Thus, PCBs represent the most likely class of anthropogenic contaminants that could negatively affect the stress response in stingrays. Temperature was the only major environmental variable to differ at the two sites, besides contamination, but was different only where females were sampled (Chapter 2). While the Q10 ratio (a measure of thermal sensitivity) is not known in this species, the average difference in water temperature between the Seal Beach Wildlife Refuge and Catalina Harbor during the sampling period was only 2.61 ± 1.07°C. Water temperatures between the open ocean mainland coast and the island where males were sampled were similar. Given the rapid nature of the acute stress response and the short duration of sampling (see below), any effect of water temperature is likely small and was not considered as a factor in this analysis. 4.2.2 Sampling. All stingrays were sampled by active hook-and-line fishing and subjected to one of two treatments: baseline or stressed. To reduce the time between capture and sampling, once a stingray in the baseline group was hooked, it was reeled in immediately, euthanized with

67 overdose of tricaine, and tissues were quickly collected and frozen for later analyses (< 2 mins from hooking to euthanasia). In this way, I was able to obtain as close to baseline levels as possible for wild caught animals. For stingrays in the stressed group, animals were reeled in slowly once hooked, giving the animal time to fight while on the line (~ 5 mins). Stingrays in the stressed group were then placed in a bucket filled with seawater and underwent a confinement stressor for 15 mins prior to euthanasia. All activities were approved by the University of Calgary under protocol AC#14-0016. Female stingrays were sampled across the breeding season from June to September (~6 per site per month) and as part of a larger study (this thesis) and contaminant-exposed males in July in 2014. Since no reference males were obtained in 2014, they were sampled the following year in August of 2015. Given the time sensitivity of the stress response, all tissues were collected in the field and were frozen on dry ice until they could be transferred to a -80°C freezer for long-term storage. Blood was collected via cardiac puncture and transferred to heparinized tubes where it remained on ice until returning to the SharkLab at California State University Long Beach where it was centrifuged and plasma subsequently frozen. After blood collection in the field, livers were removed, wrapped in foil and frozen on dry ice. A small sample (~3 g) of muscle tissue was excised from the left pectoral fin on the ventral side, wrapped in foil, and frozen. 4.2.3 Assays. 4.2.3.1 Primary response: relative 1α-OH-corticosterone A previous study demonstrated the validity of a corticosterone enzyme-linked immunoassay (ELISA) as a relative measure of 1α-OH-corticosterone concentrations in elasmobranchs by exploiting the cross-reactivity of the corticosterone antibody with 1α-OH- corticosterone (Evans et al. 2010). Since reference standards for 1α-OH-corticosterone are currently unavailable, the same corticosterone kit (Cayman Chemical, Ann Arbor, MI, item # 500655) as validated by Evans et al. (2010) was used to obtain a relative measurement of 1α- OH-corticosterone levels in individuals. This approach will not deliver absolute values of 1α- OH-corticosterone, and is sensitive to other compounds, including other steroid hormones, that might cross-react to the antibody. However, the applied stressor was expected to yield a relative increase in concentration determined by the assay Contributions from other steroids were broadly excluded by mass spectrometry. At limits of detection of 0.1 or 0.05 ng/ml by liquid chromatography coupled to tandem mass

68 spectrometry, plasma concentrations of several other corticosteroids that might cross-react with the antibody as per the kit pamphlet (corticosterone (100%), 11-dehydrocorticosterone (11%), progesterone (0.31%), cortisol (0.17%), cortisone (< 0.01%)) were rarely detectable in Round Stingrays (28/420 = 6.7% of quantitations for 103 plasma samples x five steroids)(Chapter 3). In particular, corticosterone, which has the greatest cross-reactivity with the kit antibody, was not detected by mass spectrometry in any of our stingray plasma samples. If corticosterone were present in our samples, but below the detection limit of our mass spectrometer, the most cross- reactivity that corticosterone could contribute would be 50 pg/mL (i.e. the limit of detection). This would maximally contribute less than 20% of the mean concentrations from our samples. Thus, we used the corticosterone ELISA as a way to test the change in relative 1α-OH- corticosterone measurements across the imposed stressor. For these reasons, results of the ELISA are referred to as relative 1α-OH-corticosterone concentrations. As an additional validation, concentrations of unextracted plasma (n = 2) responded linearly to serial dilution (1:4, 1:10, 1:20, 1:100: p = 0.01 and 0.006; r2 = 0.97 and 0.99, respectively; mean ± SD coefficient of variation between replicates: 1.38 ± 0.79%). Sample preparation with diethyl ether extraction was also explored, and yielded lower 1α-OH- corticosterone concentrations than unextracted plasma (n = 8, ~40% reduction) with no parallelism across serial dilution. Therefore, relative 1α-OH-corticosterone was quantified using unextracted plasma (1:2 dilution) as per kit instructions. Five percent of samples fell below the linear portion of the curve (> 80% B/Bo; three baseline stingrays, one stressed stingray). These samples were rounded up to the lowest calibrator before analyses, overestimating their relative 1α-OH-corticosterone concentrations. 4.2.3.2 Secondary response: energy substrates. Glucose, lactate and total protein content were measured in liver and muscle tissue homogenates as a proxy for tissue quality. Muscle samples were ground to a fine powder by mortar and pestle on dry ice, while subsamples of liver for homogenization were obtained by excising a small piece while the tissue remained on dry ice. A subsample of muscle or liver tissue from each stingray was added to an ice-cold 50 mM Tris buffer (pH 7.5) containing protease inhibitor cocktail (Roche, Basel, Switzerland) and sonicated on ice. Homogenates were then centrifuged at 5,000 rpm for 2 mins at 4°C in order to separate the lipid layer, prior to freezing (-80°C) aliquots of the supernatant for later analysis.

Glucose was measured using a calorimetric method (Bergmeyer, 1983) in liver and muscle homogenates and samples of plasma after protein had been precipitated out of the sample with 35% perchloric acid treatment. For liver and muscle, both free glucose and glucose stored as glycogen were quantified, whereas only free glucose was measured in plasma. Glucose stored as glycogen was quantified after treatment with amyloglucosidase following procedures of Vijayan et al. (2006). Lactate was measured in muscle homogenates and plasma enzymatically (Bergmeyer, 1983), and protein content in all samples was measured using the bicinchoninic acid method (Smith et al., 1985). Muscle and liver glucose and lactate concentrations were standardized to wet weight tissue mass. Ketone bodies (acetoacetic acid, AcAc, and 3- hydroxybutyric acid, BOH) in plasma were assayed using an EnzyChrom™ Ketone Body Assay Kit (BioAssay Systems, Hayward, CA). 4.2.4 Data analysis. Physiological parameters measured in plasma, liver, and muscle were compared within sex by site and stress state (i.e. baseline or stressed) using a multivariate analysis of variance (MANOVA). Significant dependent variables were further analyzed using a two-way ANOVA to determine the effect of site and/or stress state on individual parameter values with Bonferroni correction. All parameters were transformed (natural log) for statistical comparison, unless otherwise stated, and analyses were performed in R (v2.13.0; R Core Development Team, 2011).

4.3 Results Approximately equal numbers of individual females were sampled from the contaminant- exposed and reference site for baseline (n = 15 and 16, respectively) and stressed (n = 12 and 11, respectively) groups. Females were sampled throughout gestation from post-ovulation till late gestation. Contaminant-exposed males were effectively sampled for baseline (n = 8) and stressed groups (n = 11). However, at the reference site, male sample sizes were small (n= 5 stressed and 2 baseline). Due to the limited sample size for males, I was unable to include sex in the models and results for each sex are reported separately. 4.3.1 Females. 4.3.1.1 Baseline measures. All parameters were compared between baseline females to determine if inherent differences existed between site as well as the effect of pregnancy prior to the application of our

70 imposed stressor. Muscle lactate and plasma 1α-OH-corticosterone were the only two variables not affected by site or pregnancy (p ³ 0.24). With respect to plasma, glucose concentrations significantly increased with pregnancy (p = 0.014), but were not affected by site or an interaction (p ³ 0.23). In contrast, plasma lactate decreased over pregnancy (p = 0.007), with reference females having significantly higher concentrations during early pregnancy than contaminant- exposed females (i.e. elevated y-intercept; p = 0.011). Both liver and muscle quality were affected by pregnancy; however, there were significant interactions with site. Liver glucose content increased with pregnancy (p = 0.004), with reference females having significantly greater contents (i.e. elevated y-intercept; p = 0.031). Liver glycogen decreased with pregnancy (p = 0.03), although a weak decline was only observed in contaminant-exposed females (p = 0.044) with no change in reference females (p = 0.23). Muscle glucose and glycogen significantly decreased with pregnancy in contaminant-exposed females (p £ 0.018), while no changes were found in reference females (p ³ 0.17). The MANOVA revealed that stress had a significant effect (p = 0.002), along with site (p = 0.047), pregnancy (p < 0.0001) and an interaction between site and pregnancy (p = 0.048). Therefore, ANCOVAs and two-way ANOVAs were used to further investigate the direction of response to stress and any effects related to females sampled from contaminated versus uncontaminated sites, accounting for the effect of pregnancy. Other variables tested showed no response to capture stress or site (plasma ketone bodies, liver glycogen, muscle glycogen and muscle lactate, all p ≥ 0.08). 4.3.1.2 Primary stress response. Relative 1α-OH-corticosterone concentrations significantly increased with the application of a capture stressor in both contaminant-exposed and reference females (p = 0.042; Figure 4.2) with no effect of site (p = 0.59), and no interaction between site and stress treatment (p = 0.6). Pooling females by experimental treatment demonstrated that stressed females had significantly higher concentrations of 1α-OH-corticosterone compared to baseline females (595 ± 427 versus

413 ± 303 pg/mL; t52 = 2.11, p = 0.039). Therefore, females from both the contaminant-exposed and reference site mounted a primary stress response within the ~20 min interval from hooking to sampling.

4.3.1.3 Secondary stress response. Stress had a weak effect on plasma glucose (p = 0.056), with a significant effect of site (p = 0.008) and interaction between stress and site (p = 0.037). Among reference females, stress had the significant effect of increasing plasma glucose (p = 0.0027), despite concomitant changes with pregnancy (above). However, no change was found in contaminant-exposed females after exposure to the capture stressor (p = 0.9; Figure 4.2). Thus, contaminant exposure attenuated the glucose secondary stress response. While higher baseline values in reference females produced a significant site effect for plasma lactate (p = 0.025), there was no evidence of systematic difference in the stress applied as values in females from both sites showed predicted increases with stress (p < 0.0001; Figure 4.2), with no interactions (p ≥ 0.23).

Figure 4.2. Maternal plasma responses to capture stress by site. Plasma measures in females from our contaminated site and our reference site exposed to two different treatments. Baseline values are depicted in open bars, while stressed values in grey bars. (A) Relative 1α-OH-corticosterone; (B) plasma glucose; and (C) plasma lactate. Crosses indicate within group means and different letters indicate significant differences between groups. See text for details.

4.3.1.4 Tissue quality. Free liver glucose was the only tissue variable affected by our stress treatment, with all other measures (liver glycogen, muscle lactate, muscle glycogen) remaining unaffected (all p ≥ 0.25). In liver tissue, stress had a significant effect on free liver glucose (p = 0.023); however, closer inspection revealed that stress had no effect on free liver glucose at either site (p ≥ 0.09), although content was higher in stressed reference females than stressed contaminant-exposed females (p = 0.007). As expected, free glucose in muscle significantly decreased (p = 0.014) in response to stress as the glucose was consumed by the increased stress demands. No effect of effect of site (p = 0.68) or no stress-site interaction term (p = 0.5) was found for muscle glucose. Among the paired tissues samples from each individual, the expected inverse correlation between plasma lactate and muscle glucose concentrations was not found (p = 0.7). 4.3.2 Males versus Females. To determine the effect of sex, I compared males and females sampled within a 40-day window from each site. This window, from June 28th to August 7th, yielded a sample of 43 stingrays, distributed as 5 reference males and 7 reference females (stressed only) and 14 contaminated males (8 stressed, 6 baseline) and 17 contaminated females (7 stressed, 10 baseline). Outside of this window, pregnancy had a significant effect on tissue parameters, making them incompatible for comparison to males. 4.3.2.1 Sex effects within the contaminant-exposed site. As I had too few baseline reference males, I was only able to assess stress response differences between sexes in contaminant-exposed stingrays. However, baseline results for contaminant-exposed and reference stingrays are included in Table 4.1 for disclosure. The MANOVA revealed that both sex and stress had significant effects (p = 0.005 and p = 0.015, respectively) without an interaction. Plasma glucose, 1α-OH-corticosterone, liver glycogen, and muscle lactate, free glucose and glycogen showed no effect of stress (all p > 0.4). Most also showed no differences between males and females (all p > 0.08); however, baseline plasma glucose was significantly elevated in males compared to females (p = 0.013) as was liver glycogen (p = 0.037).

Table 4.1. Baseline male stingray plasma stress variable values. Baseline values for measured parameters between contaminant-exposed males (n = 8) and reference males (n = 2). Contaminant-exposed Reference 1α-OH-Corticosterone 321 ± 155 pg/mL 165- 195 pg/mL Plasma glucose 1.42 ± 0.56 mM 1.12 – 1.56 mM Plasma lactate 1.04 ± 0.42 mM 1.27 – 1.52 mM Liver glucose 0.23 ± 0.053 mg 0.52 - 1.59 mg glucose/g tissue glucose/g tissue Liver glycogen 10.98 ± 5.42 mg 1.66 – 2.98 mg glucose/g tissue glucose/g tissue Muscle lactate 2.20 ± 0.13 mg 1.45 – 1.64 mg lactate/g tissue lactate/g tissue Muscle glucose 0.42 ± 0.27 mg 0.25 – 0.29 mg glucose/g tissue glucose/g tissue Muscle glycogen 2.94 ± 0.84 mg 2.36 – 5.04 mg glucose/g tissue glucose/g tissue

Plasma lactate was the only parameter to demonstrate stress-related increases (p = 0.0003) that did not differ between the sexes (p = 0.3) with no sex-stress response interaction (p = 0.4). Free liver glucose content also significantly increased with acute stress (p = 0.002) and had an effect of sex (p < 0.0001), leading to an interaction between stress treatment and sex (p = 0.03). Stressed males had significantly higher concentrations than both baseline and stressed female groups (p = 0.001), with baseline males falling intermediate but significantly lower than stressed males (p = 0.001) 4.3.2.2 Site effects within the stress response between sites.

For the stressed individuals, no effect of sex, site or their interaction were found for plasma lactate, 1α-OH-corticosterone, muscle lactate, or muscle glycogen (all p ≥ 0.3; Table 4.2). Plasma glucose, liver glucose, and liver glycogen all showed significant effects of sex and site (all p ≤ 0.029), but no interaction (p ≥ 0.5; Table 4.2). Muscle glucose showed weak effects of sex (p = 0.06) and site (p = 0.08); however, posthoc testing revealed no differences among groups (p ≥ 0.6). As predicted, plasma glucose was lower in stressed contaminant-exposed stingrays than in the stressed reference population (sexes pooled, t25 = 4.8, p < 0.0001; Figure 4.3). Therefore, I found a significant effect of site on the secondary stress response that was independent of sex,

75 with contaminant-exposed stingrays less able to mobilize glucose into the blood after exposure to an acute stressor. With respect to liver glucose within site, stressed males had greater concentrations of liver glucose compared to stressed females (p ≤0.008), with stressed reference males having the highest concentrations of any group (p ≤ 0.039; Figure 4.3). Liver glycogen showed similar sex- related patterns with stressed male glycogen concentrations generally greater than stressed females; however, there was a significant effect of site (Figure 4.3). Among reference stingrays, no differences were found between males and females for liver glycogen (p = 0.9), while contaminant-exposed males had weakly higher liver glycogen contents than females (p = 0.056). Differences between males and females, regardless of site, suggest that liver carbohydrate responsiveness to acute stressors is influenced by sex, especially when females are pregnant (Table 4.2).

Figure 4.3. Sex and site effects on plasma and liver energy reserves. Sex comparisons within the stressed populations at the two sites (contaminant-exposed = dark grey bars, reference =light grey bars): plasma glucose (A), liver glucose (B) and liver glycogen (C). Crosses indicate within group means and different letters indicate significant differences between sex pairs within site. See text for details.

Table 4.2. Summary of sex differences in the response to capture stress. Comparison of blood and tissue parameters between stressed contaminant-exposed and reference male and female stingrays. Within each site, males and females were compared for significant difference and the direction of the difference is indicated. Sex-site interactions were determined when a difference existed between males and females that was not similar between sites. No difference is indicated with a dash and significant differences with a Y. Contaminant-exposed Reference Interaction Sex-effect Difference Sex-effect Difference 1α-OH------Corticosterone Plasma glucose Y Males higher - - Y Plasma lactate - - - - - Liver glucose Y Males higher Y Males higher - Liver glycogen Y Males higher - - Y Muscle lactate - - - - - Muscle glucose - - - - - Muscle glycogen - - Y Males higher Y

4.4 Discussion Anthropogenic contaminants impair the ability of teleost fishes to respond to both acute and chronic stressors (Aluru et al., 2005; Aluru and Vijayan, 2006; Hontela et al., 1992). PCBs, in particular, have drawn much attention due to their pervasiveness throughout the world and wide-reaching negative, physiological effects (Wania and Mackay, 1996). In teleosts, these contaminants have had hierarchical effects from impaired glucocorticoid signaling in the brain (Aluru et al., 2004) to disruption of steroidogenesis in the interrenal glands (Aluru et al., 2005; Aluru and Vijayan, 2006) to altering liver metabolism (Hontela et al., 1995; Wiseman and Vijayan, 2011). With respect to acute stressors, the end result of any of these PCB-related disruptions reduces or prevents the release of energy for animals to effectively remove themselves from the stressor or mitigate internal stressor effects. Pregnancy had a significant effect on modulating the stress response. When all females were compared, stingrays at both sites were able to mount a primary stress response; however, by narrowing our sample size window for comparison to males, I found stress had no effect on elevating relative concentrations of 1α-OH-corticosterone, suggesting that pregnancy may influence the primary stress response, despite it not appearing as a significant factor in the model. Another potential effect of pregnancy was the responsiveness of the liver to corticosteroid stimulation. Regardless of site, males had stronger increases in free liver glucose than females after stress. Considering the demands of pregnancy, females may be energetically compromised with respect to glucose energy stores. However, these compromises had a site-specific effect. Contaminant-exposed females experienced significant decreases in both liver glycogen stores and muscle glucose and muscle glycogen, whereas females from the reference site did not. Therefore, response to stress may be influenced by pregnancy, and its effect may be exacerbated by contaminant exposure as more energetic demands are placed upon females. While I have indications that the primary stress response is affected, results should be interpreted with caution as site differences were not strong. One explanation could be that the allotment of time between stressor application and tissue sampling was too short to capture the peak primary stress response. Species have shown wide variability the amount of time between the application of a stressor and the ability to detect physiological changes resulting from that stressor (Breuner et al., 2008; Cockrem, 2013). As females from both sites and reference males showed increases in their relative 1α-OH-corticosterone concentrations following stress

79 treatment, it is possible that differences would have increased if a longer time was allotted between capture and tissue sampling. However, allowing more time would have potentially caused a stress alteration in other physiological measures over which the baseline and stressed groups were intended to be combined (Chapters 5,6). It should also be acknowledged that our values of relative 1α-OH-corticosterone were higher than what has previously been reported in the literature for elasmobranchs (reviewed in Anderson, 2012). While matrix effects or dilution error may have affected the absolute numbers, we assume these affects to apply to all samples since they were treated the same way. Since we are interested in only relative changes and not the absolute number, differences between baseline and stressed conditions, as well as site, are reflective of underlying differences. While the liver tissue quantified for PCBs in a subset of females (Chapter 2) demonstrated significant differences by site, it should be acknowledged that females exhibited a wide range PCB concentrations, and could have influenced the magnitude of an individual’s response to stress. In particular, one female from the contaminant-exposed group had overlapping concentrations with our reference site females (715 ng/g lipid weight [lw] versus 817 ± 1,020 ng/g lw) and one reference female was within range of the mean of contaminant- exposed females (3,210 ng/g lw versus 2,856 ± 3,353 ng/g lw). In fact, PCB concentrations in the contaminant-exposed group differed over several orders of magnitude (715 – 10,900 ng/g lw) and were probably associated with age. Large differences in PCB accumulation could result in a range of responses, with more contaminated females exhibiting greater negative effects than less contaminated females. Unfortunately, I did not have PCB information for all females and was unable to relate PCB concentrations to the magnitude of the stress response for plasma variables. However, this highlights the need for field-based studies to consider that individuals vary in their environmental organochlorine exposures when interpreting results. Despite the potential confounding factor of PCB accumulation variability, I was still able to detect site-related differences in the acute stress response, suggesting stingray stress physiology is sensitive to contaminant exposure. Fishes exposed to chronic stress have shown a reduced ability to respond to punctuated acute stressors (Barcellos et al., 1999), although this phenomenon appears to be influenced by multiple factors such as species, type of stress, age and environmental factors (Barton, 2002; Davis et al., 1984; Moore and Jessop, 2003). While field-based studies do not have the benefit of

80 the history of previous stress experienced by wild caught animals, I found little indications that female stingrays were chronically stressed. Save for lactate, all baseline parameters were similar between contaminant-exposed and reference females. The higher lactate in reference females could be related to habitat accessibility by humans and/or degree of open ocean exposure. Contaminant-exposed females were sampled from a federally protected estuary that has limited public access and is sheltered from wave action. Reference females, while sampled from the back bay of Catalina Harbor, may be less sheltered and required to move around more frequently leading to an increase in baseline plasma lactate. Nevertheless, when exposed to our acute stressor, reference females demonstrated they were capable of responding to acute stress. Baseline comparisons for males between sites were more difficult due to the limited number of baseline reference males (i.e. n = 2). While there were indications that contaminant- exposed males may be chronically stressed based on differences in 1α-OH-corticosterone (Table 4.1), baseline values for contaminant-exposed males were not different from female stingrays from the same locality. It should be noted that male stingrays were sampled from a beach where it is common to see people fishing from the shore and nearby piers. Thus, fishing-related stressors may be greater for contaminant-exposed males compared to reference males that were sampled from an area with lower human fishing pressure. While it is impossible to know what, if any, chronic stressors were imposed on these wild caught stingrays, I cannot discount this as a factor that could have dampened relative 1α-OH-corticosterone responses in our contaminant- exposed male stingrays. However, strong and significant increases in liver glucose after stress in contaminant-exposed males indicates they are able to respond to acute stressors. While the degree of impact from contaminant exposure on the primary stress response in stingrays was unclear, there was a strong effect of site on the secondary stress response. For reference stingrays, both males and females showed similar and significant increases in plasma glucose after stress, whereas no response was observed in contaminant-exposed stingrays. Despite the fact that contaminant-exposed females showed indications of mounting a primary stress response, and liver glucose concentrations increased after stress, this change was not accompanied by an increase in plasma glucose as was observed in reference stingrays. The stark difference in plasma glucose response between sites indicates that stingrays from the contaminated site have a reduced ability to respond secondarily to acute stressors. However, it would be interesting to further explore whether this secondary response was completely

81 abolished or only severely attenuated by measuring stress responses over a longer time period. Our results do suggest that plasma glucose is a sensitive end-point indicator of impairment and a tool that could be extended to other elasmobranchs as an indicator of contaminant-related effects. While animal health could be a consideration for the significant difference in plasma glucose I observed between sites, I found no evidence, from a carbohydrate perspective, that contaminant-exposed stingrays had lower glucose stores than reference stingrays. In fact, I found contaminant-exposed males to have the highest hepatic concentrations of glycogen, but were unable to mobilize this energy substrate into plasma for use to respond to acute stress. Therefore, the lack of glucose response to stress in contaminant-exposed animals seems less related to tissue quality and more to disruptions in the ability to release glucose into the blood upon detection of a stressor. One mechanism by which this could occur is through disruption of glucose transporter 2 (GLUT2) gene expression, the primary glucose transporter in the liver. In mammals, administration of TCDD (another potent organochlorine with similar toxicity to dioxin-like PCBs) has been shown to significantly reduce GLUT2 expression in the pancreas, leading to impairment of glucose homeostasis and ultimately causing diabetes in animals (De Tata, 2014). TCDD is thought to suppress GLUT2 expression through the activation of the aryl hydrocarbon receptor (AhR; Matsumura, 1995; Tonack et al., 2007), which is also the route by which PCBs exert their effects (Safe et al., 1985). Other studies have demonstrated that anthropogenic chemical mixtures, in the form of treated municipal waste water effluent, can cause a decrease in GLUT2 expression in Rainbow Trout (Ings et al., 2011), suggesting that the expression of these genes in fish is sensitive to anthropogenic chemicals. Since GLUT2 is a bidirectional transporter, reduced expression in stingray liver could lead to a reduced ability of PCB-exposed stingrays to release glucose into the blood upon stress; however, further exploration of this hypothesis is required. In contrast to differences in hepatic carbohydrate release with stress between the sites, muscle function demonstrated no differences between sites or sexes. Upon hooking capture, stingrays fight to escape by vigorously flapping their wings in short bursts. As most muscle found in the wings of round stingrays is white muscle (K Lyons, pers. obser.), much of the metabolic capacity of this tissue is assumed to be anaerobic. Therefore, it was not surprising that a significant amount of lactate would be produced and, after a short period of time, measured in the blood of stingrays subjected to our acute stressor. The similar levels of lactate in plasma as

82 well as no differences in muscle free glucose or glycogen content suggests that there is no impact of anthropogenic contaminants on muscle anaerobic functioning or tissue quality, with respect to carbohydrates. Considering that the liver is the main site of organic contaminant accumulation in elasmobranchs, it is not surprising that I found no differences in muscle between sites. 4.4.1 Conclusions While Round Stingrays are mesopredators at best, even at this lower trophic position, environmental contaminants have a significant effect on their ability to mount a robust stress response and could be exerting chronic stress effects as well. Other, higher trophic level elasmobranchs are likely at risk of these impacts considering that they tend to have PCB concentrations that are magnitudes greater than levels measured in this study (Fisk et al., 2002; Lyons et al., 2013). An impaired ability to respond to stress is important considering that these larger elasmobranchs tend to also be caught in commercial and recreational fisheries. Often, management encourages a “catch and release” policy for recreational fishers and the commercial bycatch of unwanted elasmobranchs (Camhi et al., 2009; Clarke, 2013). However, if certain species, which are more prone to accumulate contaminants, have a significantly reduced ability to respond to acute stressors, such as fishing pressure, this could have important implications for fisheries management. As it stands, potential impacts of contaminants on elasmobranch survivability are not taken into consideration under current management practices. The results of this study indicate that it may be beneficial for management to consider contaminant influences on the ability of elasmobranchs to respond to stress and its potential impacts on post-capture survivability, especially for species commonly caught in recreational and commercial fisheries.

To determine whether environmental PCB contaminant exposure impaired their ability to osmoregulate, this study characterized osmoregulation in pregnant Round Stingrays (Urobatis halleri) and their embryos over the course of pregnancy. As part of a comprehensive physiological comparison between two sites with known PCB contamination differences, solutes (urea, TMAO, proteins) were quantified in matched pairs of maternal plasma and uterine fluid and activities of enzymes indirectly and directly related to urea synthesis were measured in maternal and embryonic liver tissue. Pregnant reference females maintained stable plasma urea concentrations, whereas plasma urea declined over the course of pregnancy in contaminant- exposed females. In addition, muscle protein content significantly declined in contaminant- exposed, but not reference, females, indicating a potential loss of substrate for urea formation. Embryonic enzymes involved in the urea cycle and protein processing were functional, in contrast to the hypothesis that internal gestation (matrotrophic histotrophy), would delay the developmental onset of embryonic osmoregulation. While embryos were able to maintain urea and TMAO concentrations comparable to reference embryos, their liver protein content also significantly decreased over development, suggesting that osmoregulatory costs were higher. Increased costs for osmoregulation join other physiological measures adversely affected in these stingrays.

5.1 Introduction In contrast to most bony fishes, chondrichthyans employ a unique strategy to cope with the challenges of osmoregulating in a marine environment. Chondrichthyans, including elasmobranchs (sharks, skates and rays) produce and maintain high levels of urea as their nitrogenous waste product and remain hyperosmotic to their environment (Evans et al., 2004; Smith, 1936), rather than excreting ammonium and remaining hyposmotic as is the case with marine teleosts. However, retained urea as an osmolyte has the negative effect of destabilizing proteins by altering tertiary structure and, thus, protein function (Rajagopalan et al., 1961).

Elasmobranchs circumvent this challenge by also maintaining higher levels of trimethylamine-N- oxide (TMAO) relative to teleost fishes (Evans et al., 2004) in an optimal ratio of 2:1 urea to methylamines (Yancey and Somero, 1980). Elasmobranchs produce urea through activities of the ornithine-urea cycle (OUC; Goldstein and Forster, 1971), and activities may be adjusted as elasmobranchs are exposed to different salinities (Anderson et al., 2005; Goldstein and Forster, 1971). In contrast to urea, some elasmobranchs are not able to synthesize their own TMAO (Goldstein et al., 1967; Treberg and Driedzic, 2002), while others can (Goldstein and Funkhouser, 1972; Read, 1968). Regardless of synthesis ability, TMAO in marine elasmobranchs can be obtained through their proteinaceous diets (Goldstein et al., 1967), and is excreted at low rates to reduce losses, even when the individual is energetically compromised as a result of starvation (Treberg and Driedzic, 2006). Two key OUC enzymes used by elasmobranchs are glutamine synthetase (GSase) and arginase. Unlike mammals that also produce urea as a waste product, elasmobranchs use the amino acid glutamine as the nitrogen donor, rather than NH3, to synthesize urea with GSase catalyzing the production of glutamine that is then fed into the urea cycle (Anderson, 2001, 1986). Arginase represents the final enzyme in the cycle to produce urea, and uses arginine as a substrate created either upstream in the urea cycle or directly from the diet (Armour et al., 1993). Other enzymes that have been suggested to play ancillary roles in feeding glutamate into the ornithine-urea cycle to produce the required glutamine are alanine and aspartate aminotransferases and glutamate dehydrogenase (Ballantyne, 1997). Reproductive strategy may also play a role in the timing of embryo osmoregulation development (Kormanik, 1993, 1992) as the mother contributes to regulation of the uterine environment in viviparous species (Kormanik, 1992; Thorson and Gerst, 1972). However, Read (1968b) demonstrated that embryos from a viviparous species (Pacific Spiny Dogfish Squalus suckleyi) exhibited OUC enzyme activity despite removal from the uterus, suggesting that the osmotic environment was at least partially under embryonic control. In an egg-laying species, Big Skate (Raja binoculata), embryos have hepatic arginase activity that increases during early development, with no activity documented in the yolk sac, suggesting that urea production was exclusively embryo driven. To date, embryo osmoregulation research has primarily focused on placental (Bull Shark Carcharhinus leucas) and ovoviviparous shark species (Spiny Dogfish Squalus acanthias). Histotrophic ray species with internal gestation, but no specialized placental

85 tissue, represent an intermediate level of physiological connection between mother and fetus that has not been studied. Anthropogenic contaminants are also likely to affect osmoregulation during development and in adults. Female elasmobranchs can offload organic contaminants to offspring (Lyons and Adams, 2015; Lyons and Lowe, 2013b). Given that embryonic development is a vulnerable period in an animal’s life (Elonen et al., 1998; Örn et al., 1998), contaminant exposure in utero could interfere with embryo osmoregulation. In teleosts, organic contaminants such as polychlorinated biphenyls (PCBs) and dichlorodiphenyltrichloroethane (DDT) disrupt normal functioning of gill Na+/K+ ATPase pumps (Evans, 1987). While elasmobranchs rely less heavily on their gills for ion balance than teleosts, gill cell structure modifications unique to elasmobranchs play an important role in urea retention (Fines et al., 2001), and disruptions could lead to effects on osmoregulation. Furthermore, urea synthesis for the purpose of osmoregulation is energetically expensive for elasmobranchs (Ballantyne, 1997). Therefore, metabolic costs associated with detoxification demands could reduce the pool of energy elasmobranchs have at their disposal for osmoregulation, particularly during energy intensive life stages such as pregnancy. Round Stingrays (Urobatis halleri) are abundant and easy to sample throughout their reproductive cycle (Lyons and Lowe, 2013a; Mull et al., 2010) and their embryos gestate internally, supplied with nourishment via maternal uterine secretions (matrotrophic histotrophy). In addition, adverse physiological effects on maternal and embryonic Round Stingrays are established in response to PCB contamination (Frantz, 2014; Sawyna et al., 2017; this thesis). Therefore, the objective of this study was to use blood, liver, muscle, gill tissue, kidney, and histotroph samples across pregnancy from the larger study to test the hypothesis that organochlorine contaminant exposure negatively impacts osmoregulation in mothers and embryos.

5.2 Methods 5.2.1 Field Sites Pregnant female stingrays were captured from a reference site (Santa Catalina Island, located 35 km from the coast) and a PCB-exposed site (mainland California) as part of a larger study to examine physiological responses of stingrays to environmental contaminant exposure

(this thesis). Contaminant-exposed females were sampled from within salt marshes located at the Seal Beach National Wildlife Refuge (SBNWR, 33.731N, 118.064W) and reference females from the back bay of Catalina Harbor (33.434N, 118.503W). A subset of females was analyzed for a suite of contaminants, verifying that stingrays from the mainland had significantly higher concentrations of accumulate contaminants, with PCBs contributing the most to total load (≥ 75%; Chapter 2). Wastewater treatment plants servicing the mainland’s urban centers release treated wastewater through subsurface pipes ~ 8 km offshore in waters > 200 m deep4. Due to stingray ecology, their exposure to wastewater pharmaceuticals is low to undetectable (Lyons, unpublished data), making PCBs the dominant contaminant. Females were sampled over the course of their pregnancy from post-ovulatory to near parturition, which occurs during the dry season in southern California where there is little to no rainfall. Among southern California wetlands, most are salt marshes and receive very little fresh water inputs as rain is scarce during the summer and major rivers have been altered to release their outputs directly into the ocean as a result of heavy urbanization along the coast of California (Wood, 2000). The SBNWR is tidally inundated, with salinities nearly identical to seawater (Merkel and Henderson, 2014), and Catalina Harbor is an open bay. Therefore, stingrays were exposed to at or near 100% seawater in the field with no differences in salinity between sites, removing environmental salinity differences as a major factor influencing osmoregulatory capability. Round Stingrays are not euryhaline, and their ability to tolerate small changes in salinity is not known; however, they would be expected to seek out environments with ideal salinities (i.e. 100% seawater). Temperature affects metabolism (Schmidt-Nielsen, 1997) and water temperature at the contaminated, mainland site were higher than at the reference, island site (Chapter 2). However, individuals also have the option of moving to optimal local water temperatures, so the impact is unknown. To control for temperature, all laboratory assasys were conducted at the same temperature, and litter development was aligned by a developmental marker (Chapter 2). 5.2.2 Sampling Stingrays were captured using hook and line and were subjected to one of two treatments for inclusion in a concurrent study examining the stress response: immediate euthanasia or

4 https://www.ocsd.com/residents/current-construction/outfall-land-sections-oobs-piping-rehabilitation-project-in- the-city-of-huntington-beach

87 euthanasia after a 15 mins confinement stressor (Chapter 4). Stingrays were euthanized with an overdose of tricaine (MS-222) in accordance with animal care protocols (AC#14-0016). After which, females were measured (disk width) and blood samples were obtained via cardiac puncture, transferred to a heparinized tube, and stored on ice until components could be separated via centrifugation at the lab. Plasma was then stored at -80°C until analysis. After blood sampling, the skin over the ventral side of the body was carefully removed and internal organs (i.e. liver and digestive system) were excised to expose both uteri. A small slit was made in the left uterine wall and a 10 mL syringe with an 18 gauge needled was used to aspirate histotroph. After transfer to a 15 mL tube, histotroph was immediately frozen on dry ice and subsequently stored at -80°C until analysis. Embryos were then removed, individually wrapped in aluminum foil, and frozen at -80°C for later dissection, where disk width and body, liver and kidney mass was recorded. Maternal dressed body mass (without internal viscera) and liver mass were recorded in the lab after returning from the field. Dressed mass was used to remove variability added from stomach or intestine fullness and variability contributed by other organ masses, leaving muscle as the major contributor to dressed mass. 5.2.3 Assays. Tissue homogenates were made from ~80 mg of muscle and liver from each mother and one of her embryos. Muscle tissue (30% red, 70% white in adults, 100% white in embryos; K. Lyons pers. obser.) was excised from the ventral side of the left pectoral fin and ground to a fine powder by mortar and pestle on dry ice prior to adding it to a homogenization buffer. Subsamples of liver were directly added to the homogenization buffer and manually abraded. The homogenization buffer consisted of an ice-cold solution of 50 mM Tris buffer (pH 7.5) containing a protease inhibitor cocktail (Roche, Basel, Switzerland). Tissues in buffer were then sonicated on ice for ~30s, centrifuged at 5,000 rpm for 2 mins to separate the lipid layer.

Resulting supernatant was either added to a storage buffer (21 mM Na2HPO4, 0.5 mM EDTA, 0.2% bovine serum albumin, and 5 mM β-mercaptoethanol, 50% [v/v] glycerol, pH adjusted to 7.4 ; Vijayan et al., 1997) and stored at -20°C for future enzyme activity analysis or frozen in aliquots (-80°C) for future urea and TMAO analysis. Protein content of the supernatant was measured using the bicinchoninic acid method (Smith et al., 1985). Activities were measured using the method of Mommsen and Walsh (1991) for the following enzymes in liver homogenate samples at room temperature (~23°C): alanine

88 aminotransferase (EC 2.6.1.2), arginase (EC 3.5.3.1), aspartate aminotransferase (EC 2.6.1.1), glutamate dehydrogenase (EC 1.4.1.3), and glutamine synthetase (EC 3.6.1.2). Glutamate dehydrogenase, aspartate aminotransferase, alanine aminotransferase activities were measured in all liver homogenate samples stored in glycerol buffer (-20°C), and activities were measured within four days. Activities were directly calculated from the slope of the linear portion of the consumption of reduced cofactor (at 340 nm) on a plate reader (ThermoMax, Molecular Devices, Inc) using accompanying SoftMax Pro 6 software and reported as µmoles of substrate consumed or product liberated per minute per g of liver protein. With respect to enzymes of the urea cycle, glutamine synthetase represents the first step and arginase the last step. Glutamine synthetase and arginase activities were measured using frozen homogenate (-80°C) from embryo liver samples (n = 40) and a subset of female samples (n = 19). Activities were calculated colorimetrically as product produced over a specific interval of time against a standard curve (urea for arginase and glutamyl γ-hydroxamate for GSase; Mommsen and Walsh, 1991). Activity is reported as µmoles of product produced per minute per g of liver mass (wet weight). Na+/K+ ATPase activity was measured in gill tissue homogenates from the embryos. Briefly, gill tissue was extracted from partially defrosted embryos and ground to a fine powder using mortar and pestle on dry ice. Gill powders were then homogenized and sonicated in a 50 mM Tris buffer containing a protease inhibitor cocktail (Roche, Basel, Switzerland) and enzyme activity was directly assayed following previously published procedures (McCormick, 1993). Enzyme activity is reported as µmol ADP produced/mg protein/h. Urea was measured in each pair of plasma and histotroph samples using methods outlined by Rahmatullah and Boyde (1980). Prior to analysis, each sample was deproteinated by adding trichloroacetic acid (TCA, 5% final concentration). A subsample of this solution was then diluted with milliQ water and 3 mL of the chromogenic reagent was added to 100 uL of diluted sample. Samples were boiled in a water bath for 5 mins, cooled, and the absorbance was measured at 525 nm with a urea standard curve (BioShop Canada, Inc) run in parallel with samples using a 96- well plate reader (ThermoMax, Molecular Devices, Inc). Similar procedures were followed for muscle homogenates, and urea was calculated on a per milligram tissue basis, corrected for the mass of muscle powder and volume of homogenization buffer.

TMAO was measured in plasma and histotroph samples as well as muscle homogenates following the methods of Treberg et al. (2006), scaled down by one-quarter to conserve reagents. Samples were added to 9 volumes of ice cold TCA and left on ice for 10 min before centrifugation (15.6 g for 5 min) to separate precipitated proteins. The supernatant was removed and subsequently assayed by placing TCA extracts (300 µL) into a 2 mL tube followed by 300 µL toluene and 300 µL of the iron-EDTA mixture described by Wekell and Barnett (1991). Samples were capped, heated at 50°C for 5 min, and cooled to room temperature before adding 600 µL of 45% KOH. Samples were vortexed for 15 s three times, allowing time in between mixing for separation of layers. The toluene phase was removed and added to a tube with sodium sulfate (~10 mg) before a solution of 0.02% picric acid in toluene was added. Samples were mixed and allowed to settle for 2 min before the samples were read at 410 nm on a SmartSpec Plus spectrophotometer (Bio-Rad) using a quartz cuvette. A standard curve

(TMAO•2H2O, Sigma Aldrich) was also run in parallel to samples. Osmolality of plasma and histotroph samples were measured using a Precision Systems Inc. Osmometer (model number 5004). 5.2.4 Data analysis. To correct for the reproductive asynchrony between sites (Chapter 2), clasper days (positive and negative numbers relative to the developmental stage when external sexual differentiation is first seen) were used to align litters and developmental stage (post-ovulatory, early-, mid-, or late-term) was used to distinguish yolk-dependency from histotroph nutrition (Chapter 3). MANOVAs were used to identify variables that significantly changed with site or pregnancy. The acute, experimental, capture stressor of 15 minutes in the larger study did not show any main effects in MANOVA for osmoregulatory measures (all p ≥ 0.3). Therefore, stress state was not incorporated as a factor in further analyses. Parameters were compared individually using an analysis of covariance to test for site differences. Paired t-tests were used to test for effects when tissues were taken from the same pregnant female. Pearson correlations were used to identify linear associations. All analyses were performed using R statistical package (V 2.13.0) (Team, 2011) with α set to 0.05. 5.3 Results Plasma from pregnant females was obtained in equal numbers at both sites (n = 28); however, due to six spontaneous histotroph expulsions after euthanasia but before histotroph

90 sampling, histotroph was collected from 26 contaminant-exposed and 24 reference females. Livers were excised from one embryo starting at clasper day 0 (total mass ~3g) for reference (n = 21) and Mainland (n = 19) litters. Before clasper day 0, embryonic liver could not reliably be separated from other tissues. 5.3.1 Tissue condition. 5.3.1.1 Mothers. Maternal liver protein and muscle water content were not affected by pregnancy (p ≥ 0.07). Liver water content increased across pregnancy (p ≤ 0.011) but was not affected by site (p = 0.5). On the other hand, relative liver mass (% liver mass of dressed body mass) and muscle protein content significantly decreased across pregnancy (p ≤ 0.018). For both parameters contaminant-exposed females had greater rates of decrease than reference females (i.e. steeper slope; p ≤ 0.009; Figure 5.1). This was not due to simple starvation, as all maternal stomachs showed recent feeding. However, no information was collected about prey species and quality. 5.3.1.2 Embryos. Tissue water content in embryos was not affected by contaminant exposure. Liver water content was not affected by pregnancy or site (p ≥ 0.06). Muscle water content decreased with development (p ≤ 0.002) and also did not have an effect of site (p ≥ 0.2). Liver protein content showed an interaction between site and pregnancy. There was no change over development in embryos from the reference site (p = 0.07), while liver protein content decreased in contaminant- exposed embryos (p <0.0001). Embryo hepatosomatic indices (% liver of total body mass) and muscle protein content increased with development at both sites (p ≤ 0.0003); however, embryos from the reference litters had higher values for both metrics than embryos from the contaminant-exposed litters (i.e. elevated y-intercept; p ≤ 0.038; Figure 5.1). Relative kidney mass (% of total mass) decreased with development in both sites (p ≤ 0.012), with contaminant-exposed embryos having relatively heavier kidneys than reference embryos (i.e. elevated y-intercept; p < 0.0001).

Figure 5.1. Maternal and embryo muscle condition throughout pregnancy. (A) Muscle protein content significantly declined in mothers from both the reference (circles) and contaminated site (triangles) over pregnancy using clasper days as a proxy; however, declines were steeper in contaminant-exposed females. (B) In contrast, embryo muscle protein content significantly increased at similar rates (i.e. similar slope) in reference and contaminant- exposed samples over development, with reference embryos having higher protein content (i.e. elevated y-intercept).

5.3.2 Fluid osmolytes. 5.3.2.1 Maternal plasma. Plasma proteins and osmolality were not affected by contaminant exposure or pregnancy (p ≥ 0.4). Plasma urea was significantly affected by pregnancy and showed an interaction with site. In contaminant-exposed females, urea plasma concentrations significantly declined with pregnancy (p < 0.0001), whereas no changes were found for reference females (p = 0.8). TMAO and urea:TMAO ratios only had a significant effect of site, with no changes over pregnancy. Reference females had higher concentrations of TMAO (p = 0.026), which had a corresponding effect of significantly lowering urea:TMAO ratios compared to contaminant-exposed females (p = 0.021). 5.3.2.2 Histotroph. The histotroph environment varied with pregnancy (i.e. clasper day), contaminant exposure (site) and their interaction (all p ≤ 0.0001; Figure 5.2). Over development, I predicted histotroph urea to increase as embryos excrete waste as a result of growth-related protein metabolism. Reference urea histotroph supported this hypothesis as significant increases occurred over development (p = 0.0002); however, no changes were found in contaminant- exposed samples (p = 0.5), suggesting that contaminants may negatively affect embryo urea production and/or excretion If maternal regulation of the uterine environment were occurring, I would predict TMAO concentrations to change in the same direction and magnitude as urea. Contrary to our hypothesis, TMAO concentrations did not change with development in reference samples (p = 0.4), and significantly decreased in contaminant-exposed samples (p < 0.0001). As a result, urea:TMAO ratios significantly increased with development in contaminant-exposed samples (p < 0.0001), while reference samples were able to maintain constant ratios (p = 0.1). By mid- pregnancy (i.e. clasper day 20), all contaminant-exposed urea:TMAO values exceeded the largest ratio value measured in reference samples. The significant decrease in TMAO for contaminant- exposed samples had a cascading effect on histotroph osmolality, which also decreased over development in contaminant-exposed samples (p = 0.008), but not reference samples (p = 0.59). TMAO plays an important role in counteracting the negative effects of urea (Rajagopalan et al., 1961; Yancey and Somero, 1980), and decreases in its concentration that lead to increases

93 urea:TMAO ratios are expected to have negative effects on protein stabilization and, therefore, proper protein function. Mothers are responsible for the energetic content of histotroph, which supplements embryo growth. Since embryo growth in utero is exponential (Lyons and Lowe, 2013a), mothers may not be able to keep up with embryo demands, leading to decreases in embryo number or quality. At both sites, I found significant decreases in histotroph protein concentrations over development (p < 0.0001). While histotroph from contaminant-exposed females had higher protein content during early pregnancy (i.e. elevated intercept, p = 0.022), they also experienced faster rates of decline (i.e. steeper slope, p = 0.008) than reference samples. Multiple tissues may be catabolized to create histotroph substrates, particularly proteins. Tissue type had a significant effect on the correlation between histotroph proteins and tissue proteins. No associations were found with liver proteins at either site (p ≥ 0.6), while significant correlations were found for muscle, suggesting that muscle tissue is the main source of histotroph proteins. However, contaminant-exposed females had a stronger association between muscle proteins and histotroph proteins (p = 0.015, r = 0.46), than reference females (p = 0.07, r = 0.38). This suggests that contaminant exposure compounds strains upon maternal resource provisioning to embryos. Unlike other elasmobranch species that have shown congruency between osmolyte concentrations in maternal plasma and uterine fluid (Thorson and Gerst, 1972), maternal plasma and histotroph compartments for urea, TMAO, total proteins, and osmolality showed no association, regardless of site, as developmental changes in histotroph were not reflected in plasma (all p ≥ 0.3 in all cases). Mean urea and TMAO concentrations were higher in histotroph than plasma for both sites (all p < 0.0001). In contrast, protein concentrations and osmolality were significantly lower in histotroph than plasma for both sites (all p ≤ 0.03). The lack of association between these two tissues suggests that maternal control over the histotroph osmotic environment is limited.

Figure 5.2. Maternal histotroph osmolytes. Osmolyte parameters in histotroph were compared between reference (circles) and contaminant- exposed (triangles) litters for urea (A), TMAO (B), protein content (C), and osmolality (D). Significant relationships with development are displayed with solid lines (light grey = reference, dark grey = contaminant-exposed). Mean values of female plasma parameters are depicted with horizontal, dashed lines for comparison. Asterisks indicates where post-ovulatory females were sampled (i.e. ovulated eggs with no visible embryos).

5.3.3 Muscle Osmolytes. Urea and TMAO were measured in a subset of muscle of samples from mothers (n = 13), ranging from early to late pregnancy at both sites. These osmolytes were also measured in embryo muscles starting from clasper day 0 (total mass ~3g), where sufficient tissue could be extracted (reference n =20; contaminant-exposed n = 17). 5.3.3.1 Mothers. Muscle TMAO and urea:TMAO ratios were unaffected by pregnancy (p ≥ 0.2) or site (p ≥ 0.15), with no interaction (p ≥ 0.5). However, muscle urea concentrations (nmol/mg tissue) were significantly affected by pregnancy (p = 0.02) with no main effect of site (p = 0.15) and a weak interaction between pregnancy and site (p = 0.07). Urea concentrations did not change with pregnancy in reference females (p = 0.8), but significantly decreased in contaminant- exposed females (p = 0.018). Therefore, contaminant exposure affected not only female’s ability to maintain plasma urea, but also muscle concentrations. 5.3.3.2 Embryos. Muscle urea concentrations remained unchanged with development or site (p ≥ 0.22; Figure 5.3). In contrast, muscle TMAO was affected by both development (p = 0.004) and site (p = 0.02) with no interaction (p = 0.9). Muscle TMAO concentrations significantly increased in contaminant-exposed embryos (p = 0.007) and weakly increased in reference embryos (p = 0.055; Figure 5.3). As a result, urea:TMAO ratios significantly decreased with development (p = 0.0002), with no effect of site (p = 0.09) or their interaction (p = 0.44). While embryo urea:TMAO ratios were much higher than adults during early development (4.05 ± 0.55 versus 1.93 ± 0.34), they reached ratios closer to adult values and the desired 2:1 ratio (Yancey and Somero, 1980) by near-parturition size (2.77 ± 0.37 versus 1.93 ± 0.16; Figure 5.3). Embryo muscle urea and TMAO concentrations were not correlated with histotroph concentrations for reference (p = 0.7 and p = 0.2) or contaminant-exposed samples (p = 0.5 and p = 0.08), suggesting that embryos were able to self-regulate osmolyte levels independently of changes occurring in the surrounding histotroph.

Figure 5.3. Embryo muscle osmolytes. Muscle urea (A) and muscle TMAO (B) was measured in one embryo from each litter over development from reference (circles) and contaminant-exposed (triangles) samples, with respect to mean adult muscle concentrations (horizontal dashed line). (C) Muscle Urea:TMAO ratios significantly decreased with embryo development, although they were higher than the optimal 2:1 ratio in adults (Yancey and Somero, 1980), which was replicated in this study (dashed line for mean adult female ratio). Significant relationships (p < 0.05) are shown in solid lines and weak relationships (i.e. p = 0.055) in dashed lines.

5.3.4 Urea metabolism. 5.3.4.1 Mothers. The results of our MANOVA revealed an effect of site (contaminant exposure) on enzyme activities (p = 0.044), with no effect of pregnancy (p = 0.15) or interaction (p = 0.30). Activities of liver enzymes that may metabolize proteins (i.e. amino acids) to feed glutamine into the urea cycle (glutamate dehydrogenase, aspartate and alanine aminotransferase) were not different between contaminant-exposed and reference females (p ≥ 0.12), suggesting that liver protein catabolism is not affected by site in adult females. However, enzymes directly involved in the urea cycle (i.e. glutamine synthetase and arginase) were significantly different between sites (p ≤ 0.016). As glutamine synthetase is the first step in the urea cycle for elasmobranchs (Anderson, 2001) and arginase is the last step, activities of these two enzymes are predicted to have a significant effect on urea production. Reference females had weakly higher glutamine synthetase activities than contaminant-exposed females (p = 0.08), as well as weak positive correlations with plasma urea concentrations (p = 0.076). Glutamine synthetase activity was not correlated with plasma urea in contaminant-exposed females (p = 0.23). Therefore, contaminant exposure may limit urea production by reducing the rate at which glutamine (nitrogen donor for urea) can be fed into the urea cycle. Arginase activities were also significantly higher in reference females (p = 0.03), suggesting that contaminant exposure limits urea production capability in contaminant-exposed females. However, unexpectedly, no correlations were found between plasma urea and arginase activity in reference females, likely due to the fact that plasma urea was relatively stable over pregnancy (p = 0.22). Contaminant-exposed females exhibited weak negative correlations between arginase activity and plasma urea (p = 0.065 r = -0.64). Lower enzyme activities and negative association of plasma urea with arginase activity suggests that urea production may be compromised in contaminant-exposed females. 5.3.4.2 Embryos. Development and site, as well as their interaction, had significant effects on embryo enzyme activities (all p ≤ 0.016; Figure 5.4). These developmental changes also lead to differences in activities between embryos and their mothers (Table 5.1; Figure 5.5).

Figure 5.4. Embryo enzyme activities over ontogeny. Relationships between liver enzyme activity and ontogeny for one embryo from each litter of reference (triangles) and contaminant-exposed (circles) samples are depicted. Glutamate dehydrogenase (A), aspartate aminotransferase (B), and glutamine synthetase (C) all decreased with development, whereas alanine aminotransferase (D) and arginase (E) activities increased. Gill Na+/K+ ATPase activity (F) weakly decreased with development at both sites. Significant linear regressions are displayed with solid lines and weak (p ~0.05) relationships in dashed lines.

Table 5.1. Enzyme activities across pregnancy in mothers and embryos.

Activity (mean ± SD) of embryo and maternal liver enzymes are reported and results of mother-embryo pairwise comparisons for mid- and late-term litters. Significance and direction of difference between mothers (M) and embryos (E) at these two stages is given. Embryo Mother Mid-term Mid-term Late-term Late-term activity activity significance direction significance direction Reference

Alanine 32.0 ± 10 59.2 ± 19.2 t7 = 3.73, M > E t11= 6.53 M > E aminotransferase p = 0.007 p < 0.0001 Arginase 23.7 ± 4.49 30.4 ± 4.1 NA (n = 2) M ~ E NA (n = 2) M ~ E Aspartate 29.8 ± 13.2 31.1 ± 2.88 t7 = 3.39 E > M t11 = 3.62 M > E aminotransferase p = 0.012 p = 0.004 Glutamate 8.71 ± 4.60 6.17 ± 2.40 t7 = 5.33 E > M t11 = 2.15, M ~ E dehydrogenase p = 0.001 p = 0.054 Glutamine 3.91 ± 2.03 61.8 ± 21.6 NA (n = 2) M > E NA (n = 3) M > E synthetase Exposed Alanine 30.3 ± 17 80.7 ± 12.5 t5 = 5.16, M > E t12 = 3.93, M > E aminotransferase p = 0.004 p = 0.002 Arginase 25.7 ± 6.97 29.3 ± 7.53 NA (n = 1) M ~ E NA (n = 2) M ~ E Aspartate 18.3 ± 12.8 49.1 ± 18.6 p = 0.42 M = E t12= 4.60, M > E aminotransferase p = 0.0006 Glutamate 4.84 ± 2.22 9.93 ± 5.47 p = 0.7 M = E p = 0.26 M = E dehydrogenase Glutamine 3.75 ± 3.62 35.9 ± 13.9 NA (n = 1) M > E NA (n = 2) M > E synthetase

100

Figure 5.5. Maternal and embryonic enzyme activity comparisons. Glutamate dehydrogenase (A,B) and aspartate aminotransferase (C,D) activity varied between mother (dark greys) and embryo (light grey) pairs depending on embryo developmental stage (Chapter 3) and by site. Reference stingrays (top panels) showed more differences between embryos and mothers in terms of activity by stage compared to mothers and embryos from the contaminant-exposed site (bottom panels). Different letters denote significant differences in pairwise comparisons within developmental stage.

101

Enzymes involved with protein metabolism showed an interaction with pregnancy and site. Glutamate dehydrogenase activity declined at both sites (p < 0.0001), although rate of decline was weakly stronger in reference embryos than contaminant-exposed embryos (p = 0.062). Aspartate aminotransferase activities had a significant interaction with pregnancy as activities declined in reference embryos (p < 0.0001), but not in contaminant-exposed embryos (p = 0.13). On the other hand, alanine aminotransferase activities significantly increased with development at both sites (p <0.0001); however, site had a significant effect on both the slope (p = 0.044) and y-intercept (p = 0.004), with contaminant-exposed embryos having a greater rate of activity increase with development than reference embryos. Since three enzymes involved with protein processing changed in different directions over development and by site, I was interested to determine the association between enzyme activity and tissue quality. Aspartate aminotransferase activity was not correlated with liver protein content at either site (p ≥ 0.12). Glutamate dehydrogenase had a positive correlation with liver protein content at both sites (p ≤ 0.008, r ≥ 0.57). In contrast, alanine aminotransferase had a negative correlation with protein content in contaminant-exposed embryos (p = 0.001, r = - 0.77), but no relationship in reference embryos (p = 0.16, r = -0.33). Therefore, total protein catabolism was calculated as the difference between alanine aminotransferase and glutamate dehydrogenase activity. In contaminant-exposed embryos, protein catabolism was negatively correlated with liver protein content (p < 0.0001, r = -0.81), while this relationship was weaker in reference embryos (p = 0.057, r = -0.43; Figure 5.6). Since protein catabolism increases with development (p ≤ 0.01), the stronger negative association found in contaminant-exposed embryos may indicate that contaminants negatively impact protein content in exposed embryos while reference embryos were better able to spare their liver proteins over development. This negative effect on proteins could have implications for growth and tissue quality (Chapter 2,6). Interestingly, I found opposite effects of development on urea cycle enzymes and an interaction with site. Glutamine synthetase activity declined at both sites (p ≤ 0.0001), and while the rate of decline was similar (p = 0.15), contaminant-exposed embryos had higher activities (i.e. elevated y-intercept) than reference embryos (p = 0.020). The ontogenetic decline in glutamine synthetase activity may act to spare proteins for growth once embryos have established their osmotic regulatory system. In contrast, arginase activity, the last step producing

102 urea, significantly increased with development both sites (p ≤ 0.0001). Developmental changes in arginase activity were similar between sites (i.e. similar slopes, p = 0.85); however, reference embryos had higher activities than contaminant-exposed embryos (i.e. elevated y-intercept, p = 0.004). Arginase activity had a weakly positive association between arginase activity and histotroph urea in reference samples (p = 0.084, r = 0.4), but no association in contaminant- exposed samples (p = 0.13, r = -0.39). This suggests that arginase activity in reference embryo results in the production of urea (excreted to the histotroph), but disruption of urea production in contaminant-exposed embryos despite having the enzymatic capability to do so.

103

Figure 5.6. Accelerated protein catabolism in contaminant-exposed embryos Negative associations were found between embryo liver proteins and protein catabolism (defined here as the difference in alanine aminotransferase and glutamate dehydrogenase activity) for embryos at the reference (circles) and contaminant-exposed site (triangles). Given that protein catabolic activity increases with development (see text), the stronger association found in contaminant-exposed embryos has implications for protein processing and growth.

104

5.4 Discussion 5.4.1 Plasma vs. histotroph. Given the potential uterine surface area for material exchange, it was surprising that histotroph did not more closely reflect concentrations found in maternal plasma. Studies in species with less surface area (i.e. placental or ovoviviparous species) found concentrations of solutes in these two compartments to be closely correlated, leading authors to suggest that the histotrophic environment was controlled by the mother. The findings of our study suggest that mother’s control of the uterine environment is influenced by several factors including diffusion and transport dynamics and possibly reproductive mode. Furthermore, our study indicates that embryos do have osmoregulatory capability early in development and should not be discounted as playing a role in influencing their uterine environment as implied in other studies. Lack of correspondence between these two compartments could result from differences in their physiological roles for maternal and embryo compartments and/or or chemical differences in osmotic concentration gradients influencing solute exchange. As an exclusively marine species, Round Stingrays are expected to maintain osmolalities higher than saltwater

(~1000 mOsm/kg H2O). In the current study, I found that female plasma had osmolalites close to that of seawater, while osmolalities of uterine fluid were significantly lower. Differences between plasma and histotroph might be due to differences in osmoregulatory demands, whereby mothers need to maintain higher plasma osmolalities for their own osmoregulatory functioning. Maintaining the uterine environment hypoosmotic to maternal plasma must require energy, suggesting lower histotroph osmolality is adaptive. On the other hand, solute exchange is dictated not only by its own concentration gradient but also the gradients of other solutes, which all contribute to the overall osmotic gradient. The difference between maternal plasma osmolality and urea concentration was much greater than that in histotroph, suggesting that other solutes in maternal plasma contribute to osmotic pressure, whereas in histotroph urea appears to be the major contributor. Because maternal plasma osmolality tended to be higher than histotroph, this could hinder the rapid exchange of urea. Therefore, urea may not be exchanged from histotroph (higher concentration) to plasma (lower concentration) because it would be acting against the osmotic pressure gradient, which was higher in maternal plasma and lower in histotroph. For euryhaline species where maternal

105 plasma osmotic pressure can be altered by movement into different environments, lowering maternal plasma osmotic pressure may facilitate the exchange of urea from the histotroph to maternal plasma, leading these tissues to be correlated more closely. This hypothesis is supported by work in Bull Sharks, where the histotrophic environment closely resembled maternal plasma (Thorson and Gerst, 1972). Unlike Bull Sharks, Round Stingrays are stenohaline and the lack of ability to enter brackish or freshwater may constrain maternal ability to use environmental salinity to moderate osmotic balance. Despite the higher urea concentrations in uterine fluid compared to plasma, I did not see evidence of uterine flushing, as has been proposed to occur in other species. Even when stressed (Chapter 4), uterine conditions were similar compared to unstressed females, indicating that females are resistant to flushing and only relax sphincters in life-threatening situations to abort a pregnancy. Instead, the uterine environment isolates embryos from the external environment. Histotroph is energetically expensive, and thus uterine flushing would be a large waste of resources. The use of uterine flushing as an osmoregulatory strategy is likely to be dependent on matrotrophic investment, and I would predict that females with high investment would be less prone to flushing. Thus, flushing in lecithotrophic species like the Spiny Dogfish (Kormanik, 1992) may occur due to the lack of maternal input beyond egg resources, but not occur in species like the Round Stingray where supplemental maternal investment is high. 5.4.2 Embryo osmoregulatory capability. Despite the fact Round Stingray embryos develop internally, creating the opportunity for their mothers to manage osmoregulatory challenges, embryos gain the capacity to osmoregulate at a young age. Previous studies have speculated that a benefit of internal gestation is to delay development of embryo osmoregulation systems in order to lower embryo energetic demands by sparing energy that can be put instead into growth. However, the present study documented that not only are key OUC enzymes present early in development, but they experience developmental changes throughout ontogeny as well, suggesting they are functional. However, this study quantified enzyme activities in optimal conditions and, thus, is likely to underestimate physiological limitations in vivo. Activity is a proxy for amount of enzyme available in tissues, but the actual amount of product created is likely to be constrained by substrate availability. In addition, other factors such as water temperature, time since feeding, and nutritional state could also influence enzyme activities.

106

While maternal capacity will always be physically greater than embryos due to their significantly larger liver mass, activity rates per gram tissue of the two enzymes were similar between embryos and their mothers. In particular, arginase activity, the enzyme responsible for the last step in urea synthesis, was not different between mother-embryo pairs. Despite the increase in embryo arginase activity over development from both sites, the activity of glutamine synthetase, one of the first enzymes in the urea cycle, decreased with development, such that late term embryos had lower activities than their mother’s enzyme activities. Correspondingly, enzymes that feed glutamate into the urea cycle (i.e. aspartate aminotransferase and glutamate dehydrogenase) decreased with development as well and had activities that were lower than adults. This decrease of upstream enzyme activities could be responsible for the weak correlation between embryo arginase activity and histotroph urea concentrations (reference site only), and a mechanism by which embryos spare protein for growth rather than urea production. Embryos maintain an osmotic boundary between themselves and the uterine environment (i.e. histotroph). Using muscle as a whole-body proxy, urea muscle concentrations showed no change with time, despite the finding that urea histotroph in reference samples increased with development. However, embryo muscle urea was also significantly greater than adults, suggesting that embryos are not completely impermeable to urea uptake. Due to size constraints, I was unable to obtain embryo plasma samples, but future studies comparing maternal, uterine and embryo compartments may help elucidate solute transfer and osmoregulatory abilities of embryos. Interestingly, I found indications that embryos in an internally gestating species are capable of retaining and concentrating TMAO and this ability develops as they growth, similar to that documented in skate embryos developing outside their mother’s body in egg cases (Read, 1968). Currently, the ability to synthesize TMAO appears to be species specific (Goldstein et al., 1967; Goldstein and Funkhouser, 1972; Treberg and Driedzic, 2006) and how this changes with development is unknown. The lack of correlation between muscle and histotroph TMAO concentrations would suggest that embryos can synthesize TMAO; however, this needs confirmation. Furthermore, despite the lower histotroph levels of TMAO in contaminant-exposed samples, the muscle tissue of embryos was comparable to reference embryos, indicating they were able to still maintain comparable levels in the face of a limited external source (i.e. histotroph). The higher muscle urea:TMAO ratio in stingray embryos contrasts with findings

107 from skate embryos, which were capable of maintaining the desired 2:1 ratio (Steele et al., 2004). One major difference between these species is their mode of reproduction as skates develop in egg cases outside of the mother’s body and are exposed to much lower levels of “environmental” urea during development. 5.4.3 Contaminant effects. While the marine environment in southern California remains highly anthropogenically influenced, not all species are exposed to the same chemical factors, based on their ecology and historic and current release of contaminants into the environment. Major chemical contributors in the area include dichlorodiphenyltrichloroethane and metabolites (DDTs) and pharmaceuticals. However, the nearshore preference of stingrays significantly limits their exposure to these other factors on the mainland, leaving PCBs as the major contaminant group known to accumulate in their livers. A deep channel (> 700 m in depth) and northward moving current prevents contaminant transfer from the mainland to the island and also functionally prevents stingrays from moving between sites. Since human activity on the island in areas where stingrays were sampled from is low, our island stingrays represent an appropriate reference site when employing a field-based study. Given that other major variables were accounted for (especially salinity), we cautiously attribute site effects to differences in maternally accumulated PCBs. With regards to negative effects in utero, muscle urea and TMAO concentration similarities between embryos from our two sites provided no support for the hypothesis that maternally offloaded contaminants significant affected the osmoregulation of embryos. However, the cost of osmoregulation may be different between sites, with more cost to contaminant-exposed embryos. In particular, liver protein content was negatively correlated with our proxy for protein catabolism in contaminant-exposed embryos, which resulted in greater decreases in liver protein content over development and, therefore, lower embryo quality. One issue with determining impacts on osmoregulation when using muscle variables is that muscle also has urea synthetic capabilities (Steele et al., 2005), in addition to liver. Thus, concentrations measured in muscles may not be reflective of other tissues. Unfortunately, due to the collecting procedure (i.e. flash freezing embryos in the field), and small size of embryos, I was not able to collect embryo plasma, which would have allowed for a more direct comparison of embryo plasma, maternal plasma and histotroph. Nevertheless, embryos were similar at the

108 contaminant-exposed and reference sites, suggesting embryos are less affected by PCB contamination than their mothers, at least in utero. Unlike embryos, I observed significant site-related differences between mothers with respect to enzyme activities and how those activities changed over pregnancy. Key enzymes in the urea cycle had higher activities in reference females, which possibly resulted in reference females being able to maintain plasma urea levels during pregnancy, whereas urea plasma concentrations fell in contaminant-exposed females. In addition, muscle quality degraded at a faster rate in contaminant-exposed females over pregnancy. Muscle represents a large cache of protein available to feed amino acid substrates into the urea cycle if protein obtained from food does not meet demands. Interestingly, despite the decrease of muscle proteins, presumably for urea production, contaminant-exposed females were not able to maintain plasma urea. Urea is not only an important osmolyte needed for maternal osmoregulation, but is also energetically expensive, so loss of this osmolyte will have energetic cost to the individual. Considering that pregnancy represents a time high energy demand energy strains for females, the added cost of dealing with contaminant exposure is likely to compromise osmoregulatory ability and/or efficiency. If impaired urea retention is responsible for the decrease in plasma urea, it suggests that a primary site of contaminant effects may be at the kidneys or gills, where mechanisms are in place to reduce urea loss. Previous studies have demonstrated that the combination of elasmobranchs’ unique cell membrane structure (high cholesterol content) and sodium-coupled, secondary active urea transporters functions to greatly reduce urea loss across the gills (Fines et al., 2001). It would be interesting to examine the effect contaminant exposure has on gill impermeability or renal reabsorption efficiencies to determine if exposure impairs stingrays’ ability to retain urea. Furthermore, exploring whether this potential indication of “osmotic regulation stress” is influenced by pregnancy status would be of value. As females have high- energy demands during pregnancy, any extra energy that goes towards dealing with contaminant exposure may tip the scale such that females are unable to osmoregulate efficiently. 5.4.4 Conclusions. The present study demonstrates that stingray embryos develop osmoregulatory capability early in life and further support the hypothesis that urea synthesis is both costly, and homeostatically important in elasmobranchs. I also found that the uterine environment was

109 influenced by the embryo, rather than being under maternal control in Round Stingrays. Given the wide range in matrotrophic strategies utilized by elasmobranchs, it will be interesting to explore how embryo osmoregulation is affected by the extent of maternal-fetal contact. Our study demonstrates that exposure to PCBs can strain maternal osmoregulatory demands and reduce efficiency. Further studies should investigate whether similar results are found in males to determine the degree that pregnancy represents general impacts of stingray osmoregulation in the face of contaminant exposure.

110

Organisms have constrained energy budgets and any additional cost, such as that added by contaminant exposure, represents energy taken away from growth or reproduction. Little is known about the sublethal effects of contaminant exposure on elasmobranch energy budgets. We compared tissue quality and metabolic capacity in male and female Round Stingrays (Urobatis halleri), as well as embryos, from two populations experiencing different levels of environmental PCB exposure. With histotrophic matrotrophy, pregnancy is energetically costly in this species. Pregnant females from the exposed population experienced larger decreases in tissue quality and mass than reference females. In addition, higher PCB burden resulted in decreases in relative hepatic energy reserves, supporting the hypothesis that contaminant exposure compounds energy demands during pregnancy. Despite the large energetic strain on pregnant females, contaminant- exposed males had the highest metabolic capacities accompanied with the lowest tissue quality metrics. We support the accumulating evident that in utero exposure to contaminants via maternal offloading results in negative embryo outcomes. Embryos from the reference site had higher tissue quality measures, with similar metabolic capacities as contaminant-exposed embryos, suggesting contaminant exposure reduced resource utilization efficiencies. These sex and age class-related impacts of contaminants are likely to have broad, negative, population-level consequences for elasmobranchs.

6.1 Introduction Metabolism is the sum of all chemical reactions that are required to sustain life. Energetic fuels (i.e. carbohydrates, lipids, proteins) are needed to supply energy to maintain these processes for survival and reproduction. Environmental toxicant exposure is a known stressor that imposes additional energetic costs, as energy must be diverted to repair the affected systems or remove the contaminant through detoxification processes (Beyers et al., 1999; Marchand et al., 2004; Smolders et al., 2005). Thus, energy devoted to dealing with contaminant exposure

111 represents energy taken away from growth, reproduction, immune function, and various behaviors, among other important functions. Little is known about how well elasmobranchs cope with environmental contaminant stressors. In Spiny Dogfish (Squalus acanthias) injected with various titers of silver nitrate and copper, survival decreased as exposure increased (De Boeck et al., 2007). However, for other contaminants, such as organic contaminants, very little is known about the lethal and sublethal toxicity of accumulation in elasmobranchs, despite their propensity to accumulate organochlorine contaminants (Fisk et al., 2002; Lyons et al., 2013; Mull et al., 2013). As part of a larger study to examine the physiological consequences of PCB contaminant exposure using the Round Stingray (Urobatis halleri) (this thesis), differences in energy reserves and metabolic capacity were compared in stingrays from our reference and contaminated sites. Unlike teleosts, elasmobranchs (sharks, skates and rays) are not able to oxidize free fatty acids (FFA) in their muscles (Ballantyne, 1997) and levels of circulating FFA in the plasma are low (Larsson and Fänge, 1977). Instead, elasmobranchs convert hepatic lipids into ketone bodies, which, along with amino acids, are consumed as the preferred fuel source (Ballantyne, 1997). Elasmobranchs not only use the liver as their primary lipid storage organ, but lipids also contribute a high proportion to total liver mass. Thus, metabolic impacts of organochlorines are expected to have an impact on liver lipid reserves. Tissue quality measurements of liver proteins (Chapter 5), carbohydrates (Chapter 4), and lipids (Chapter 2) plus liver enzymatic metabolic capacity (a battery of 8 enzymes representing key processes in metabolism) were chosen as appropriate measures to determine the metabolic burden of PCB exposure. In addition to the primary hypothesis that PCB contamination would decrease energy reserves and increase metabolic capacity, sex differences were predicted. During pregnancy, internal gestation with histotrophic matrotrophy represents a large energetic investment (Wourms and Demski, 1993), with near-term litters accounting for 15-20% of their mother’s total mass (Chapter 2). On the other hand, pregnancy also offloads contaminants onto developing embryos, which has the potential to reduce individual impacts of a contaminated environment for adult females but not males (Hickie et al., 2007; Ylitalo et al., 2005). Therefore, our study had three objectives. 1) to estimate the metabolic cost that PCB exposure adds to female energy expenditures during pregnancy; 2) to assess whether contaminant exposure affects adult males and females differently; and 3) to determine the effect of maternally offloaded contaminants on

112 embryo metabolism and liver quality.

6.2 Methods 6.2.1 Study Sites Males and females were captured at two full seawater locations around Seal Beach, California, (33.739N, 118.113N and 33.731N, 118.064W, respectively) and Santa Catalina Island (33.4427N, 118.498W and 33.434N, 118.503W), the former representing our contaminated site and the latter our reference site. A deep ravine (~700 meters) separates Santa Catalina Island from the mainland, a geological feature proposed to be responsible for the recent genetic separation between stingray populations (Plank et al., 2010). As such, organic contaminant analysis of a subset of stingrays used in this study confirmed previous work that PCBs comprise a majority of accumulated organic contaminants (≥ 75%; Chapter 2). Although it remains possible that the selection pressure of contaminants has selected for genetic change in the mainland populations, the selection interval has been short. The peak of organic contaminant production occurred in 1970, and this study was conducted only 44 years later (2014), which is equivalent to ten generations. Furthermore, based on mitochondrial DNA, elasmobranch evolution occurs at a slower rate than mammalian evolution (Martin et al., 1992). Therefore, genetic divergence is not expected to play a major role in physiological responses to PCB exposure. Besides environmental contamination, temperature was the other important variable to differ between the sites. However, temperature only differed between areas where females were sampled due to contaminant-exposed females being sampled from an insulated saltmarsh compared to females from the reference site, with no difference between areas where males were sampled. Temperature influenced the timing of ovulation, with contaminant-exposed females ovulating earlier than reference females. We accounted for this temporal difference by aligning females and embryos based on a developmental marker (see below). 6.2.2 Sampling Liver and muscle tissues from pregnant females, adult males, and one embryo from each litter were obtained from a study including an acute capture stress of 15 min duration (Chapter 4). In addition, six frozen livers (-80°C) from reference male stingrays sampled in August were obtained from a concurrent study (Sawyna et al., 2017). This yielded a total sample size of 29

113 females and 8 males from the reference site (37 total) and 30 females and 11 males from the contaminated site (41 total). After euthanasia, whole livers were collected and a sample of pectoral muscle tissue was taken. 6.2.3 Tissue quality. Tissue homogenates were first created as outlined in Chapter 4. Protein was quantified by the bicinchoninic acid method (Smith et al., 1985), lipids via the Folch method (Folch et al., 1957), and embryo carbohydrates as outlined by (Bergmeyer, 1983). Lipid was not quantified in muscle samples as they contain minimal lipid (Lyons, unpublished data). The energy content (kJ/g) for each liver fuel source was then determined using coefficients from Jonsson et al. (1997) for proteins (17 kJ/g tissue), lipids (38 kJ/g tissue), and carbohydrates (17 kJ/g tissue; embryos only). Total liver energy reserve was calculated by summing the two fuel sources (kJ/g). Muscle tissue was expressed as mg protein per gram of homogenized muscle (wet mass). An acutely applied capture stress (Chapter 4) had a significant effect on adult muscle and liver carbohydrates. As carbohydrates contribute little to the total energy content (≤ 1%) of the liver, they were excluded from adult quality analyses. Carbohydrate contributions to embryo energy balance was assessed, as the short duration of the stressor did not deplete in utero carbohydrates. 6.2.4 Metabolic capacity. For the purposes of this study, metabolic capacity was defined as the maximal activity of key enzymes from a range of metabolic processes involved with protein, carbohydrate, and lipid processing. This was quantified in all adult males and females, as well as one embryo from each litter developmentally advanced enough to allow liver sampling (> 3 g total mass = clasper day 0, n = 40). Activities were measured in liver homogenates for the following enzymes: aspartate aminotransferase (E.C. 2.6.1.1), alanine aminotransferase (E.C. 2.6.1.2), glutamate dehydrogenase (E.C. 1.4.1.2), hexokinase (E.C. 2.7.1.1), lactate dehydrogenase (E.C. 1.1.1.27), phosphoenolpyruvate carboxykinase (PEPCK; E.C. 4.1.1.32), pyruvate kinase (E.C. 2.7.1.40), and 3-hydroxybutyrate dehydrogenase (E.C. 1.1.1.30), following previously described procedures (Mommsen and Walsh, 1991; Wiseman and Vijayan, 2011). Activities were expressed as µmoles of NADH produced or consumed (measured at 340 nm) per min per gram protein. Enzymes were grouped into categories based on the energy substrate they affect: proteins (alanine and aspartate aminotransferase, glutamate dehydrogenase), carbohydrates (hexokinase, lactate dehydrogenase, PEPCK, pyruvate kinase) or ketone bodies (3-

114 hydroxybutyrate dehydrogenase). Total metabolic capacity in this study was defined as the sum of all enzyme activities (µmoles/min/g) within an individual. I also calculated substrate-specific capacity as the sum of enzyme activities within each substrate group. As expected based on the short stress duration, enzyme activities did not differ between baseline and stressed adults. Thus, stress was not included in subsequent analyses. 6.2.5 Data analysis. Given the reproductive asynchrony in our populations, pregnant females and embryos were aligned between sites based on an embryonic developmental marker (clasper days). Claspers are the male copulatory organ that develop as an extension from the distal tip of each pelvic fin. Clasper days are defined as the days relative to the appearance of external sexual differentiation of males that allows anatomical sexing of the embryo (Chapter 2). Tissue total energy content and metabolic capacity was compared between sites among adult females to examine the effect of pregnancy. Linear regressions for each tissue variable were used to calculate the change in quality between early pregnancy (clasper day 0) and the last day of embryo collection at the contaminated (82 clasper days) and reference (68 clasper days) sites. As PCBs were measured in a subset of females from each site (Chapter 2), I also examined the effect that individual contaminant burdens had on those pregnant females. I calculated the residual values using site-specific regressions of total hepatic energy (kJ/dw) against clasper day. Residual values were then compared to PCB burden. To examine the effect of sex and site, MANOVAs for adult tissue quality and enzyme activities, followed by two-way ANOVAs on parameters showing significance in the overall model, were employed. Pearson’s correlations were used to examine the response of tissue quality parameters with metabolic capacity based on substrate grouping. The effect of development on embryo metabolism (starting at clasper day 0) was explored by MANOVA, with clasper day, sex, and site as independent factors. Significant factors were further explored with linear regressions. Since females are known to maternally offload (Lyons and Lowe, 2013a), I predicted that embryos from more contaminated mothers would experience negative impacts on energy reserve acquisition during development. Embryos paired with mothers with known PCB burden were identified, and assigned their individual residual from the relationship between reference embryo total mass versus clasper day. Litter median residual values were then compared against maternal PCB concentrations.

115

As embryos draw all of their resources from their mother, I also predicted a tradeoff between embryo quality and maternal quality. To quantify this tradeoff, I compared residuals from mother-embryo pairs to examine whether higher relative embryo quality came at the expense of maternal quality. For mothers, I obtained residuals from a variety of within-site relationships between clasper days and maternal quality (i.e. relative liver mass, liver energy content (kJ/g), total liver energy (kJ/dw). I obtained residuals from relationships between liver mass and total energy (kJ), liver mass and lipid content (kJ/g), embryo total mass and liver lipid content (kJ/g). In addition, I obtained residuals from the relationship of embryo total mass against clasper day and then took the median value for each litter. A variety of combinations of maternal-embryo residual pairs were then tallied for the following categories for each site based on the sign of the residuals: positive embryo-positive mother, positive embryo-negative mother, negative embryo-positive mother, and negative embryo-negative mother. For each category, values were tabulated and a chi-square test was used to determine if distributions differed from the null (equal distributions among categories) as well as if distributions differed between sites.

6.3 Results 6.3.1 Effect of pregnancy. Pregnant females from the reference (n = 29) and contaminant-exposed site (n = 30) were sampled from June (post-ovulation/early pregnancy) through September (late pregnancy) (Chapter 2). Pregnancy had little effect on relative body mass (dressed mass/disk width) and was not affected by site (p ≥ 0.6). Dressed mass (body mass without internal viscera) was used as it removes variability added from stomach or intestine fullness or variability contributed by other organ masses as well as representing mass contributed by mostly muscle. From early to late pregnancy, females experienced small declines in body mass for contaminant-exposed (4.0%) and reference females (1.4%). In contrast, liver mass with respect to size (g liver/disk width) significantly decreased with pregnancy at both sites (p ≤ 0.013), with no difference between slopes (p = 0.09). While contaminant-exposed females had significantly heavier livers during early pregnancy (p = 0.0007), their longer pregnancy period resulted in larger decreases (~50% reduction) compared to reference females (~30% reduction; Table 6.1). The small changes in body mass, but large changes in liver mass, indicate that the liver is a sensitive indicator of pregnancy effects on body condition.

116

Table 6.1. Tissue changes across pregnancy by site. Regression coefficients for tissue variables where females showed effects of pregnancy and site. The relative difference between the beginning of pregnancy (i.e. y-intercept) and end of pregnancy represents the relative change females underwent at each site. End values were calculated from the last day of sample collection based on clasper days (i.e. 68 days for reference females and 82 days for contaminant-exposed females; Chapter 2). Tissue quality measures are values reported on kJ of substrate per gram of tissue for liver and mg of substrate per g of tissue for muscles. Difference in and direction of regressions between PCB-exposed (E) and reference (R) females are given. Slope Intercept Ending % Slope Y- (early-term Value change with site intercept value) pregnancy effect site effect Relative - E > R liver mass Reference -0.0090 2.01 1.39 30.47 Exposed -0.0170 2.77 1.38 50.24 Total liver - E > R energy Reference -0.0819 22.35 22.35 24.91 Exposed -0.0588 24.18 24.18 19.93 Liver lipids - E > R Reference -0.0834 20.30 14.63 29.60 Exposed -0.0625 22.81 17.68 22.48 Muscle E > R E > R proteins Reference -0.0455 22.0 18.96 14.0 Exposed -0.1054 27.0 18.36 32.0

To assess liver energy conversion efficiency into offspring mass, I determined the difference between each female’s theoretical liver energy (total kJ) at clasper day 0 and her liver energy on the clasper day she was sampled based on the relationship between female liver energy (total kJ/disk width) and clasper day within site (reference: kJ/DW = - 0.3212*clasperday+46.39, p = 0.0008, r2 = 0.33; exposed: kJ/DW = -0.4464*clasperday+65.05, p <0.0001, r2 = 0.45). While energy loss generally increased with pregnancy, the theoretical asymptote for reference females was lower (~500 kJ) compared to contaminant-exposed females (~770 kJ). Thus, reference females were more efficient at converting maternal resources into offspring biomass (Figure 6.1).

117

Figure 6.1. Conversion of energy reserves into litter mass. Theoretical liver energy loss (kJ) of pregnant females (difference between energy at clasper day 0 and day of sampling) with increasing litter mass. Females from the contaminant-exposed site (dark grey triangles) had greater losses than females from the reference site (light grey circles).

118

6.3.1.1 Tissue quality. Pregnancy was predicted to have a significant, negative impact on female tissue quality. As with liver mass, liver total energy content (kJ/g) significantly decreased with pregnancy (p = 0.0002), demonstrating that both liver mass and quality are negatively impacted with pregnancy. However, liver energy content also had a significant site effect (p = 0.024). While rate of decrease was similar between sites (p = 0.52), contaminant-exposed females had higher total energy contents than reference females (p = 0.023; Table 6.2). With respect to energy groups, hepatic energy content significantly declined with pregnancy (p ≤ 0.0001) and had an effect of site (p ≤ 0.02). Both contaminant-exposed (p = 0.027) and reference (p = 0.001) females showed the decline across pregnancy, with early-term contaminant-exposed females having higher liver lipid reserves than reference females (i.e. elevated y-intercept; p = 0.019). While the rate of decrease was similar between sites (i.e. similar slope, p = 0.57), reference females had larger relative declines in liver lipid content (30%) compared to contaminant-exposed females (22%; Table 6.1). No main effects or interaction effects were found for liver protein content (p ≥ 0.07). To determine the effect that PCBs had on liver quality with pregnancy, I compared total liver energy (kJ/dw) residual values (see data analysis) against PCB concentrations. PCB concentrations had a weak negative effect on liver quality (sites combined, p = 0.062, r2 = 0.23). Site and pregnancy had a significant effect on muscle tissue quality (Chapter 5). Muscle protein content decreased across pregnancy (p < 0.0001); however, decreases were not similar between sites (i.e. slope, p = 0.037). Contaminant-exposed females had both higher muscle protein content (i.e. y-intercept, p = 0.0016) and steeper slopes than reference females. Therefore, muscle protein change from early to late pregnancy was twice as large in contaminant-exposed compared to reference females (32% versus 14%; Table 6.1).

119

Table 6.2. Adult hepatic enzyme activities by site and sex. Mean ± standard deviation activities (µmoles/min/g) for our eight enzymes by site and sex. Different letters represent significant differences among sex-site groups for each enzyme. Reference Reference Exposed Exposed Females Males Females Males (n = 8) (n =8) (n =10) (n = 11) Alanine 54.73 ± 19.81a 65.99 ± 12.25ab 56.72 ± 21.04a 105.7 ± 55.76b aminotransferase Aspartate 25.55 ± 6.47a 34.50 ± 6.14ab 38.52 ± 19.01ab 57.19 ± 26.62b aminotransferase Hexokinase 3.37 ± 1.28 5.00 ± 0.95 4.20 ± 1.68 5.48 ± 2.60 Glutamate 4.94 ± 1.72a 11.46 ± 1.99bc 8.61 ± 4.95ab 14.24 ± 6.03c dehydrogenase Lactate 6.18 ± 2.64a 4.69 ± 0.813ab 8.12 ± 3.55b 8.89 ± 3.27b dehydrogenase Phosphoenolpyruvate 2.51 ± 5.44a 0.68 ± 0.78a 3.37 ± 8.89a 0.82 ± 0.61a carboxykinase Pyruvate kinase 2.92 ± 0.75a 14.99 ± 2.74b 4.88 ± 2.81a 13.02 ± 6.29b 3-hydroxybutyrate 0.31 ± 0.19a 0.41 ± 0.19a 0.57 ± 0.33a 0.36 ± 0.21a dehydrogenase

6.3.1.2 Metabolic capacity. Neither pregnancy nor site had an effect on total metabolic capacity (all p ≥ 0.27). Within substrate groups, protein metabolic capacity (glutamate dehydrogenase, aspartate and alanine aminotransferase) did not change with pregnancy and was not different between sites (p ≥ 0.23). However, there was a negative association between protein metabolic capacity and liver protein content (kJ/g; p ≤ 0.0006; r = -0.61). Carbohydrate metabolic capacity (hexokinase, lactate dehydrogenase, pyruvate kinase, and PEPCK) was affected by site (p = 0.021), but not pregnancy (p = 0.58) with no interaction (p = 0.24). Contaminant-exposed females had higher activities than reference females (p = 0.008). Like protein, carbohydrate metabolic capacity was inversely associated with carbohydrate energy content, in contaminant-exposed females (p = 0.036), although the effect was small (r = -0.38) and not seen among reference females (p = 0.33). Similar to carbohydrates, ketone metabolic capacity (i.e. 3-hydroxybutyrate dehydrogenase activity) exhibited a significant effect of site (p = 0.0003), with no effect of pregnancy or interaction (p ≥ 0.43). Contaminant-exposed females had significantly higher activities than reference females (p = 0.0002). Using a partial correlation between lipid content

120 and ketone metabolism controlling for clasper day, revealed no correlation between lipid content and ketone metabolism in females from either site (p ≥ 0.21). Thus, carbohydrate and ketone metabolic capacity were elevated at the contaminant site. Total metabolic capacity was not correlated with PCB concentration in the subset of 16 females (p = 0.46). Among substrate groups, carbohydrate capacity had a weak, positive correlation with PCB levels (p = 0.062, r = 0.46). Protein and ketone capacities exhibited no correlations (p ≥ 0.71) 6.3.2 Adult sex differences. To determine the effect of sex, I compared adult males and females sampled within a 40- day window from each site (Chapter 4). This window, from June 28th to August 7th, yielded a sample of 37 stingrays, distributed as 8 reference and 11 contaminated males, and 8 reference and 10 contaminated females. Outside of this window, representation of males and females across sites was too unbalanced for the model. 6.3.2.1 Tissue quality. Sex had a significant effect (p = 0.0013) on relative liver mass (g/disk width), with no effect of site (p = 0.85), although there was a site-sex interaction (p = 0.011). Contaminant- exposed females had significantly larger livers for their size compared to males from the same site (p = 0.0006); however, no differences were found between reference males and females (p = 1; Figure 6.2). MANOVA revealed that site (p = 0.001) and sex (p < 0.0001) had a significant effect on adult tissue parameters. Total energy content was higher in females (p < 0.0001) with no effect of site and no interaction (p ≥ 0.093; Figure 6.2). Liver protein had an effect of sex (p = 0.022) and site (p = 0.032), but no interaction (p = 0.56). Liver protein was highest in reference males, and lowest in contaminant exposed females, with reference females and exposed males intermediate (Figure 6.2). Similar to total energy content, liver lipid was higher in females than males (p < 0.0001), with no effect of site (p = 0.06) or interaction (p = 0.075; Figure 6.2). Muscle protein content was affected by sex (p = 0.025) and had a sex-site interaction (p = 0.0002). Reference females had lower protein content than males and exposed females, with exposed males intermediate.

121

Figure 6.2. Adult tissue quality measures by site and sex. Comparison of tissue quality between adult males and females for (A) relative liver mass, (B) total hepatic energy content, (C) liver proteins, (D) liver carbohydrates, (E) liver lipids, and (F) muscle proteins between reference (light grey) and contaminant-exposed site (dark grey) stingrays. Letters represent significant differences among groups and crosses indicate means.

122

6.3.2.2 Metabolic capacity. Total metabolic capacity showed a strong effect of sex (p = 0.005) and an effect of site (p = 0.03), with no interaction (p = 0.30). Contaminant-exposed males had significantly higher metabolic capacities than females from both sites with reference males intermediate (Figure 6.3). Protein capacity was significantly affected by sex (p = 0.006) and site (p = 0.026), with no interaction. Contaminant-exposed males had higher activities than reference females with reference males and exposed females intermediate (Figure 6.3). In addition, a negative association between contaminant-exposed male protein metabolic capacity and liver protein energy content (p = 0.001, r = -0.83) was not found for reference males (p = 0.6). Carbohydrate metabolic capacity was affected by sex (p = 0.009) with no effect of site or an interaction (p ≥ 0.20; Figure 6.3). However, posthoc comparisons showed that the only significant difference was contaminant-exposed males having higher capacities than reference females (p = 0.037), with contaminant-exposed females and reference males being intermediate. No significant associations were found between male carbohydrate capacity and liver carbohydrate energy content (p ≥ 0.3). Ketone metabolic capacity did not show any effect of sex (p = 0.38) or site (0.22), or interaction (p = 0.063; Figure 6.3). No associations were found between liver lipid energy content and ketone metabolic capacity (p ≥ 0.2). Individual enzyme activities can be found in Table 6.2.

123

Figure 6.3. Adult metabolic capacity by substrate group. Total metabolic capacity (A) and enzymes grouped by substrate effect were measured in adult males and females from our reference (light grey bars) and contaminated (dark grey bars) site. Protein capacity (B) represents the sum of enzyme activities for glutamate dehydrogenase, alanine aminotransferase and aspartate aminotransferase. Carbohydrate capacity (C) represents the sum of hexokinase, lactate dehydrogenase, PEPCK, and pyruvate kinase. Ketone capacity (D) is the activity of 3-hydroxybutyrate dehydrogenase. Letters denote post-hoc statistical differences among groups.

124

6.3.3 Embryo metabolism. 6.3.3.1 Tissue quality. Unlike adults that have sex-related differences in energetic demands and habitat utilization, male and female embryos have similar energetic priorities and are exposed to similar environments in utero. Therefore, it was not surprising that our MANOVA resulted in no significant effect of sex (p = 0.71) or site (p = 0.31), but a significant effect of development (p = 0.003). However, there was a significant interaction between development (clasper days) and site (p = 0.032). Relative embryo liver mass (g/disk width) increased across development at both sites (p < 0.0001), with a weak effect of site (p = 0.066) and interaction (p = 0.077), as reference embryos had slightly larger livers for their development compared to embryos from the contaminant-exposed sites (Figure 6.4). For total hepatic energy, development was the only factor positively affecting embryos (p = 0.004), with no effect of site (p = 0.96). Among substrate groups, liver carbohydrates significantly increased with development at similar rates at both sites (p < 0.0001), although reference embryos had significantly higher carbohydrate energy contents (i.e. elevated y-intercept; p = 0.031; Figure 6.4). Liver protein content significantly declined with development in embryos (p = 0.0002) with no effect of site (p = 0.58). Liver lipids were significantly affected by development (p = 0.018); however, positive increases with development were only found in contaminant-exposed embryos (p = 0.006) as no significant changes occurred in reference embryos (p = 0.49; Figure 6.4). Muscle carbohydrates significantly increased with development (p < 0.0001), with reference embryos having slightly higher quality (i.e. elevated y-intercept, p = 0.07) and rates of increase (i.e. slope, p = 0.068). On the other hand, muscle protein content was significantly affected by site (p = 0.025) along with development (p < 0.0001). Reference embryos had higher muscle protein contents than contaminant-exposed embryos.

125

Figure 6.4. Embryo quality by site. Embryos showed several site-related effects for relative liver mass (A) and tissue quality parameters (B-D) between reference (grey circles) and contaminant-exposed embryos (dark grey triangles). Since total metabolic capacity had a significant interaction with development, sex, and site, data is shown separately for males and females from each site. Significant relationships are denoted with solid lines.

126

Embryos did exhibit substantial departures from adults in terms of both hepatic energy content and composition. While energy content per gram liver in embryos was not substantially different from adults (11.3 ± 1.8 kJ/g versus 19.4 ± 6.1 kJ/g), the contribution of energy components was different (Figure 6.5). While always dominant, lipids comprised a much lower proportion in embryos (77 ± 4.2%) than in adult males (83 ± 9.8%) or females (91.3 ± 5.4%). Therefore, proportional energy contributions of proteins and carbohydrates were greater in embryos than in adults.

127

Figure 6.5. Relative hepatic energy reserves. The hepatic energy composition of stingray livers was influenced by both sex and size class. Adult stingrays (A) had higher relative proportion of lipids (light grey), while embryos (B) had greater energy contributions from proteins (medium grey) and carbohydrates (dark grey). Within group means and standard deviations (bars) are shown for each sex-site group.

128

6.3.3.2 Metabolic capacity. Unlike tissue quality, development had less effect on metabolic capacity in embryos. Total metabolic capacity had no main effects of development (p = 0.26) or embryo sex (p = 0.28), but a weak main effect of site (p = 0.076), which lead to significant interactions (p ≤ 0.042; Figure 6.4). Among reference embryos, males and females showed weak changes in total metabolic capacity with development (p = 0.074), and no effect of embryo sex (p = 0.63) or their interaction (p = 0.66). Contaminant-exposed embryos also showed no effect of development (p = 0.27); however, there was a significant effect of embryo sex (p = 0.018) and an interaction (p = 0.011). Development had a positive effect on total metabolic capacity in contaminant-exposed male embryos (p = 0.009), with no effect in contaminant-exposed female embryos (p = 0.38). In addition, contaminant-exposed male embryos were also the only sex-site group to exhibit a significant, positive correlation between total metabolic capacity and total hepatic energy content (p = 0.003, r = 0.08; all others p ≥ 0.09). As main effects, development, site and embryo sex did not influence carbohydrate metabolic capacity in embryos (p ≥ 0.11), however, there was a weak interaction between sex and site (p = 0.079). In addition, a weak, positive correlation was found between embryo carbohydrate metabolic capacity and liver carbohydrate content (all embryos combined, p = 0.067, r = 0.29). Protein metabolic capacity was significantly affected by development (p = 0.034), without main effects of sex (p = 0.29) or site (p = 0.11), but an interaction between development and site (p = 0.022). Reference embryos were found to have a negative relationship between development and protein metabolic capacity (p = 0.004), while contaminant-exposed embryos exhibited no developmental changes (p = 0.40). Ketone metabolic capacity weakly increased with development (p = 0.055), and was not affected by sex, site or their interaction (p ≥ 0.13). Similar to adults, embryos showed no correlation between lipid stores and ketone metabolic capacity (p ≥ 0.21). Maternal PCB burden was not correlated with metabolic capacity in embryos (p ≥ 0.47). 6.3.3.3 Maternal-embryo tradeoffs. As embryonic demand on their mothers increase with development, I might predict there to be a tradeoff between embryo quality and maternal quality, leading to more mother-embryo pairs falling within the “positive embryo, negative mother” category. However, within sites, I found no difference in the distribution of pairs among categories from the null (i.e. even

129 distribution; p ≥ 0.026). With only one embryo randomly sampled per litter, and no knowledge of each mother’s energy status at the beginning of pregnancy, I was unable to address embryo- maternal tradeoffs with this dataset. For samples with corresponding PCB information, most maternal-embryo pairs fell into the negative embryo residual category (7/12), with only two having marginally positive residuals for mother-embryo pairs. Therefore, there was a scarce representation of positive embryo- positive-mother pairs, suggesting negative outcomes of PCB exposure to either mothers or embryos.

6.4 Discussion With other environmental variables considered, differences in PCB accumulation remained our most important site difference separating mainland and island stingrays when comparing animals within sex groups. Otherwise, sex significantly influenced tissue quality and metabolic capacity, with site often having an interactive effect, suggesting that PCB exposure does not affect adult males and females in the same way. Despite the clear energetic burden placed on females, contaminant-exposed adult males were more energetically impacted than females. Given the high maternal reproductive investment required by matrotrophic histotrophy, this was unexpected. Future elasmobranch studies should involve both sexes to determine the generality of this finding, as elasmobranch males might be even more vulnerable to PCBs than females. 6.4.1 Effect of pregnancy. Pregnancy is an energetically demanding process and maternal investment in offspring for matrotrophic elasmobranchs is quite high (Wourms et al., 1988). Both contaminant-exposed and reference females exhibited declines in liver lipid quality and mass over the course of pregnancy, coupled with decreases in tissue protein (Chapter 5) and carbohydrate (Chapter 4) measured previously, this underscores the high energetic costs of histotrophic matrotrophy. Similar to other elasmobranchs (Abdel-Aziz and El-Nady, 1993; King, 1984), hepatic lipid content significantly declined over the course of pregnancy in females from both sites. With respect to our two main tissues of interest (i.e. liver and muscle), the liver bore the majority of the energetic burdens of pregnancy compared to muscle tissue, as dressed body mass (which is mostly muscle) showed little decrease over pregnancy. Therefore, female stingrays withdraw

130 resources to sustain pregnancy primarily from the liver. This concentration of lipids within the liver may spare or protect other tissues from catabolism during pregnancy, although other behaviors such as increased feeding may also help spare catabolism of other tissues as well. I predicted that the greatest impacts of contaminant exposure would be seen during the time when animals are most energetically strained, such as pregnancy in females. Indeed, I found that between our two sites, females from the contaminant-exposed site had the greatest decreases in relative liver mass and total body mass (dressed + liver mass), which were accompanied by decreases in liver and muscle quality compared to reference females. In support of our hypothesis, I found several indications that metabolic capacity was greater in contaminant- exposed females than reference females, suggesting that exposure increases energetic demands. In addition, contaminant-exposed females had stronger negative correlations between proteins and carbohydrates and their substrate group metabolic capacity, suggesting that when activities increase their energy reserves are significantly more impacted than at the reference site. However, the lower temperatures at the reference site could also have a protective effect for pregnant females by lowering standard metabolic rate, although this was not quantified. The negative effect of PCB accumulation on relative total hepatic energy, supports our hypothesis that increased PCB exposure exacerbates energy loss during pregnancy. One explanation for the larger decreases in tissue mass and quality in contaminant- exposed females could be greater offspring investment. On the other hand, I demonstrated that contaminant-exposed stingrays are less efficient at converting maternal energy into offspring biomass. Larger mass decreases without greater positive gains in embryo mass represents a significant cost to females from the exposed site as they will have to replenish larger energy deficits after parturition before reproducing again. Therefore, contaminant exposure leaves females energetically compromised and could have impacts on female lifetime fitness. 6.4.2 Adult sex differences. Among adults, I did observe an interactive effect of site and sex, with contaminated- males being generally more negatively affected than reference males. Not only did contaminant- exposed males tend to have lower quality measures, but they also had the highest total metabolic capacity. Higher metabolic capacity coupled with lower tissue measures suggests that contaminant-exposed males have a significant, added strain to their energy budgets, which I attribute to the cost of dealing with contaminant exposure. Increased physiological demands

131 would be expected to translate into negative outcomes for non-essential activities such as growth. Using total body mass data from previous work (Lyons et al. 2014; Sawyna et al. 2017) where comparable data was available at both sites, I found that reference males had significantly greater mass increases for their disk width compared to contaminant-exposed males (Figure 6.6). However, detailed age-growth studies would be needed to confirm this hypothesis. Reproduction, including mating and sperm production, is another area where contaminant exposure may strain energy availability for this activity. Future research should explore the consequences lower tissue quality has on contaminant-exposed male reproductive success and the potential for population-level consequences. Contaminant-exposed males had greater, negative effects than contaminant-exposed females, as pregnant females were still able to maintain mean tissue qualities and metabolic capacities comparable to reference males and females. This discrepancy, especially considering the energetic burdens placed upon females at this time, was unexpected. However, previous reports have commented on the potential sex-related differences between contaminant-exposed males and females (Lyons et al., 2014), suggesting that during pregnancy females may be able to suppress transcription and translation of enzymes involved in detoxification. In the context of our study, this could potentially represent significant energy savings. However, it would be interesting to determine how males and females compare when females are not gestating and if other ecological factors may play a role (i.e. differences in habitat, food quality, and contaminant exposure).

132

Figure 6.6. Male mass relative to disk width by site. Pooling data from the current study and that of Lyons et al. (2014) and Sawyn et al. (2017), total body mass was compared between adult males starting at the smallest size (18 cm DW) where data were available from both reference (light grey circles) and contaminant-exposed (dark grey triangles) stingrays. Mass per unit size increased at a greater rate in reference adult males (total mass = 79.2DW-1158, r2 = 0.76) compared to males from the contaminant-exposed site (total mass = 59.6DW-778, r2 = 0.64; p = 0.016).

133

Nevertheless, outside of site-related effects, there were differences that were solely attributed to sex, and likely reflect the different energetic strategies of males and females. In particular, the energetic composition of the liver was clearly skewed towards lipids in females, and more so towards proteins and carbohydrates in males, although lipids still comprised the vast majority of the energetic composition. This is unsurprising given that females were pregnant and investing energy into offspring. Males, on the other hand, may be more focused on “protecting” proteins that can be used for growth and consuming lower energy yielding substrates, such as carbohydrates, more routinely. Because adult males and females enact different strategies in terms of the energy substrates they accumulate, this could have implications for the amount of surplus energy each sex has available to cope with external or internal stressors, thereby influencing the degree to which each sex is affected. Males, by having fewer lipid stores, may be more energetically vulnerable than females, especially when females are not pregnant. 6.4.3 Embryo metabolism. Development had a significant, positive effect on tissue quality, as would be expected with rapid embryo growth. Elasmobranch reproductive strategies center upon producing few but well provisioned offspring. The production of relatively large offspring is thought to reduce predation risk of young elasmobranchs (Castro, 1983) as well as provisioning young with large amounts of energetic resources prior to birth (Hussey et al., 2010; Wourms et al., 1988). However, embryos must balance growth with storage, which results in different energetic strategies than adults as reflected in their widely different compositions of relative liver energy substrates. In addition, embryos had lower metabolic capacities than adults (101 ± 19 versus 140 ± 68 µmoles product/min/gram protein), which could represent an energy-saving technique of the liver to allow more resources for extra-hepatic tissue growth or a reflection of the underdevelopment of this organ in utero. It would be interesting to examine how metabolic capacity changes from embryonic to neonatal stages when young elasmobranchs are independent from their mothers. Embryos also differed from adults in how they responded to contaminant-exposure (Table 6.3). In adults, sex had a prominent effect both regardless of site and its interaction with site. Site had several significant interacting effects with development in embryos, with sex only appearing as a significant factor in total metabolic capacity differences. In this case, contaminant-exposed male embryos were the only group to demonstrate significant increases in

134 metabolic capacity, which could be a factor for why male embryos were not relatively heavier than females at the contaminated site as they were at the reference site (Chapter 2). The appearance of weak sex effects in utero indicates that contaminant exposure may differentially affect males more than females and that these differences intensify during adulthood, possibly at around sexual maturity when energetic priorities of males and females diverge.

Table 6.3. Main effects and interactions for site, sex, and development. Comparison of main effects of sex, site and development (in embryos only) and their interactions (denoted by “:”) for adult and embryo stingrays sampled from reference (R) and contaminant- exposed (E) sites. For embryos, the result of the significant effect(s) is reported. Adult effects Embryo effects Direction (embryos) Relative liver mass Sex, sex:site Development, weak R > E site, weak site:development Total liver energy Sex Development Increases Liver proteins Sex, site Development Decrease Liver lipids Sex Development, Increase in E site:development No change in R Liver carbohydrates Site (Chapter 4) Development, site R > E Muscle proteins Sex, sex:site Development, site R > E Muscle - Development, weak R > E carbohydrates (Chapter 4) site Total metabolic Sex, site Weak site, Increase in E males, capacity development:site, No change in E females development:site:sex No change in R embryos Protein capacity Sex, site Development, Increase in E embryos, development:site No change in R embryos Carbohydrate Sex - - capacity Ketone capacity - - -

Nevertheless, I did find site-related differences among embryos, suggesting that in utero contaminant exposure is consequential. Our hypothesis that growth was more efficient in reference embryos based on differential pregnancy lengths (Chapter 2) was confirmed by this study for these same samples. For several tissue metrics, reference embryos had higher qualities than contaminant-exposed embryos. Thus, reference embryos were more efficient at converting maternal resources into growth throughout their gestation. Site differences disappear when embryos are compared on a total mass basis (p ≥ 0.21), and site has a smaller effect on embryos than adults suggesting that females adjust their litter size to protect offspring quality. While in

135 utero, embryos must stockpile energy for use after birth when embryos become self-reliant as well as take on new energetic demands (i.e. foraging, predator avoidance, etc.). PCB burden was not quantified in embryos. The null hypothesis is that the embryonic PCB burden would be proportional to concentrations found in their mothers. However, in utero effects may not be as great as in adult females if embryo detoxification pathways were not fully developed. Specifically, the aryl hydrocarbon receptor-CYP1A system is implicated. CYP1A is an enzyme whose expression is induced through toxicant binding (i.e. dioxins, PCBs) to the aryl hydrocarbon receptor (AhR), which is implicated as the mediator of many negative effects caused by dioxin-like contaminants (Nebert et al., 2000). Preliminary data from this work found embryo hepatic expression of CYP1A to be low to undetectable (data not shown) in both contaminant-exposed and reference embryos, suggesting low biochemical responses to exposure. Embryonic detoxification pathways, including the ontogeny of CYP1A expression are therefore worthy of further investigation. 6.4.4 Conclusions. In the current study, we attribute site-related effects in tissue quality and metabolic capacity to overall differences in PCBs accumulation, with interactions found between PCBs, sex, and age class. Changes in metabolic demands over the course of an animal’s life results in not only the reshuffling of energetic priorities but also how other factors, such as PCB exposure, impact those processes. During periods of energy strain for females (i.e. pregnancy), negative effects of PCB exposure are compounded. Interactions between site and sex were most apparent in adult where males from the contaminant-exposed population were differentially affected compared to females and reference males, suggesting that males may be the more vulnerable sex to contaminant impacts. Differential effects on the fitness of males and females present a scenario where contaminant exposure has the potential for population-level implications in elasmobranchs. As contaminant effects in utero are presumed, future work should identify the critical point in development where sex-related differences become most apparent and the impact this may have on energy budgets of animals with respect to other important life processes such as growth and reproduction

136

Chapter 7: Perspectives and Future Directions

7.1 Project Summary Despite production bans nearly fifty years ago, PCBs continue to have adverse effects on wildlife health, highlighting the need to integrate toxicological assessments with management decisions. PCB contaminant impacts also do not exert their effects equally between the sexes or across the life stages of an animal. This dissertation extends our understanding of contaminant impacts on stingray physiology and toxicological biomarkers. Results suggest that contaminant exposure does affect elasmobranch physiology. Specific findings are highlighted below. 7.1.1 Reproductive impairment in females. Adult females were sampled throughout the reproductive season. I predicted that contaminant burden would increase the energy expenditure of pregnant females, resulting in poorer body condition and an impaired ability to respond to an acute stressor. Results supported this hypothesis and are described below. 7.1.1.1 Pregnancy was expensive. As expected, pregnancy with matrotrophic histotrophy in the Round Stingray is an energetically demanding process. Round Stingray mothers invest heavily into producing large and well-developed young, which presumably has the effect of increasing survival rate by being bigger (i.e. less predation risk) and more well provisioned (i.e. less chance of starving) than they would be otherwise. As the major energy storage organ of elasmobranchs, it was not surprising that the liver was taxed more heavily than the body overall (primarily muscle), despite the fact that muscle comprises a greater portion of total body mass (~ 85%, Lyons, unpublished data). However, unlike the liver, muscle requires energy to fulfill its only role, whereas energy storage and output are normal roles of the liver Not only did liver mass decrease, but liver quality decreased as well, demonstrating that overall energy loss significantly decreases with pregnancy. 7.1.1.2 Contaminant-exposure was energetically compromising. Regardless of contaminant exposure, pregnancy resulted in both mass and energy loss in females from both sites. However, overall resources were more strained in females from the contaminated site, suggesting that contaminant exposure exacerbates energy loss during pregnancy. Females from the contaminated site had faster losses of lipids (the most energy rich substrate), liver mass, and muscle proteins. These greater losses did not translate into higher

137 quality offspring, meaning females spend more energy for less (by clasper day) or equal (by embryo mass) gains than females from the reference site. Females from the contaminated site also were unable to maintain proper osmotic balance, particularly for the critical osmolyte, urea. Urea is not only a foundation of elasmobranch osmoregulation, but is energetically costly to produce. A contributing factor to the higher loss of energy in contaminant-exposed females is likely due to their inability to retain urea. Carbohydrate and ketone metabolic enzymatic capacities were also higher in contaminant-exposed females, supporting the hypothesis that contaminant exposure increases metabolic demands, which leads impairments in maternal condition at the end of pregnancy. 7.1.1.2 The acute stress response was impaired. Having established the effects of contaminant exposure on pregnancy energy balance, an acute capture stressor was introduced. Female stingrays were able to respond by mounting a primary stress response within 15 mins. However, contaminant-exposed females were unable to mount a secondary response and mobilize glucose into the blood as effectively as reference females. Instead, contaminant-exposed females only showed increases in lactate after stress, suggesting dysfunction was at the level of the liver. Impairment of the secondary stress response is expected to impact the abilities of contaminant-exposed stingrays to effectively cope with challenges, such as escaping predation. Thus, pregnancy was more energetically demanding for females from the contaminant-exposed population, affecting essential homeostatic processes from osmoregulation to liver function and stress-induced mobilization of energy substrates. 7.1.2 Adult males have impairments from contaminant exposure as well. Samples were compared across a 40 d window when females were pregnant, and males were, therefore, not seeking mating opportunities. As such, I predicted that male energy budgets and homeostasis would be less challenged than females. Results suggest the opposite. 7.1.2.1 Sex influences physiology. Unsurprisingly, sex influenced tissue quality, energy metabolism, and responsiveness to an acute stressor. One of the most striking differences between males and females was the composition of liver energy reserves. Males had lower lipid and higher protein than females. It is not known whether these differences would persist outside of pregnancy. Despite their relative inability to mobilize liver glucose into the blood, contaminant-exposed males also responded to acute stress with greater increases in free liver glucose than females, which mirrored that found

138 at the reference site, suggesting that pregnancy compromises the responsiveness of the liver to corticosteroids. In general, males also had greater metabolic capacities (enzyme activities) than females. 7.1.2.2 Contaminant exposure impaired male homeostasis. As males were not seeking mates at the time of sampling, evidence that males were more impaired by contaminant exposure than pregnant females was unexpected. Contaminant-exposed males typically had higher metabolic capacities than reference males, suggesting an elevated potential for energy generation. This higher metabolic capacity was linked to lower energy reserves and lower tissue quality. Males from the reference site had comparable liver sizes to females, while contaminated males had significantly lighter livers than their female counterparts. Because males also tend to carry fewer lipids, smaller livers exacerbated the energy content differences between males and females at the contaminated site. Therefore, males were closer to the energetic tipping point than females, even when those females were pregnant. In addition, body mass relative to size (disk width) was lower in contaminant-exposed males, suggesting that contaminant exposure diverts energy that would otherwise be contributing to somatic growth. Similar energetic compromises might affect other aspects of male life history from mate-seeking to sperm quality. Thus, not having to bear the energetic costs of histotrophic pregnancy, did not protect males from physiological consequences of environmental contaminant exposure. 7.1.3 Embryos influenced their uterine environment and competed for maternal resources. As a basic contribution to our understanding of elasmobranch pregnancy, this study showed that embryos influence their own uterine environment (i.e. histotroph). 7.1.3.1 Embryonic steroids and osmoregulatory capacity were highly developed. Embryos were capable of synthesizing steroid hormones as well as osmoregulating from very early developmental stages. This does not support the prevailing assumption that elasmobranch mothers with internal gestation control histotroph composition. Steroid hormone concentrations, particularly for progesterone and testosterone, peaked near the time of sexual differentiation, implicating embryo steroidogenesis. Embryos also expressed enzymes critical for osmorgegulatory homeostasis, and maintained histotroph osmotic environments that differed from maternal plasma. These findings, and the close maternal connection of many female elasmobranch species have with their young, suggest that embryos are important architects of their own uterine environment.

139

7.1.3.2 Embryos competed for resources. I also found evidence of resource competition within a litter. With a matrotrophic strategy such as histotrophy utilized by Round Stingrays, mothers cannot control how resources are distributed. As mothers’ resources are exhaustible, intrauterine competition is predicted. The mass of the largest embryo in a litter was not affected by litter size, suggesting that the quality of one pup is maintained through competition. Increasing mass variability among littermates with litter size and development also supports our intrauterine competition hypothesis. These findings, and the close anatomical connection between mothers and their embryos in many elasmobranch species, suggest that that Round Stingray might be a valuable model for testing the generality of mammalian maternal-embryo signaling pathways in elasmobranchs. 7.1.4 Maternal contaminant exposure affects her developing embryos. 7.1.4.1 Contaminants impaired embryonic resource utilization. When liver mass and body mass were compared between sites (i.e. in utero contaminant exposures) I found very few differences. However, when the two sites were aligned by developmental stage, results supported the hypothesis that contaminant exposure had negative impacts on embryo growth. Reference embryos were more efficient at converting maternally supplemented energy into biomass compared to their contaminant-exposed counterparts. Despite the warmer temperatures at the contaminated site, which are presumed to have a positive effect on embryos by accelerating growth, reference embryos were able to overtake contaminant- exposed embryos in terms of total body mass after one month of development. Specifically, contaminant-exposed embryos did not use their lipids efficiently, shunting lipids towards liver storage rather than somatic growth. Compared to reference embryos, this suggests dysfunction in energy utilization. Osmoregulatory capabilities did not differ at the two sites, although this finding did not rule out greater energy expenditure by contaminant-exposed embryos for this process. The production of less energetically efficient offspring suggests that decreased survival is likely. 7.1.4.2 Embryo sex matters. Contaminant exposure (i.e. site) also had an interaction with sex for embryo growth. Among dominant embryos, reference males were significantly heavier than females, whereas no sex-related differences were found in contaminant-exposed embryos. In addition, sex-related mass differences were exacerbated when litters were mostly composed of males, suggesting that

140 the sex composition of litters has a significant effect, which, based on the literature for mammals, could have lifelong impact. In addition, total metabolic capacity increased with development in contaminant-exposed male embryos, but not female or reference embryos. These sex-related differences in utero, combined with even larger differences I found between adult males and females, suggest that early life exposure to a contaminated environment is likely to reduce future survivorship and possible reproductive success. 7.1.5 Conclusions. In summary, legacy PCB contamination continues to have adverse effects on individual Round Stingray homeostasis, manifested in metabolic capacity, energy reserves, and reproductive success. Thus, elasmobranchs, even ones occupying lower trophic levels, are vulnerable to organochlorine toxicity, yet remain under-studied. 7.2 Future Directions The range of physiological impairments found in this species has large implications for other elasmobranchs. Despite the fact that Round Stingrays are not a high-performance fish, and a mesopredator at best, they still had significant PCB burdens, due to their proximity to highly industrialized and urbanized areas. Comparing PCB concentrations (both wet and lipid weight) measured in the current study to previous reports for elasmobranchs worldwide, demonstrates that stingrays are relatively contaminated (Figure 7.1). Furthermore, PCB concentrations in Round Stingrays (current study) compared to other species where contaminant effects have been investigated, revealed that Bonnethead Sharks and Atlantic Stingrays had PCB concentrations that were 40% and 25% less concentrated, respectively, than Round Stingrays (Gelsleichter et al., 2006, 2005; Manire, 2002). In those studies, reproductive effects corresponding to PCB contaminants were unclear, while in the present study I was to quantify site-specific effects. This suggests that there is a threshold contaminant level in order for effects to be observed, and that the threshold is likely to be species-specific. Therefore, future work on contaminant impacts on elasmobranchs should be extended to more species, exhibiting a range of physiological adaptations. Other factors to be considered that may influence species’ thresholds, include the importance of trophic position (as a factor contributing to exposure) and how magnitude of exposure correlates with effects. In the present study, I used site as a proxy for exposure; however, measuring contaminants in a subset of females demonstrated that individuals are exposed to a wide range of contaminants.

141

While the effect of site, as a whole, was significant, I do not know what a true baseline state looks like, as a few of our PCB subset of reference stingrays also had exposures comparable to contaminant-exposed females. Therefore, the range of potential individual exposures could have obscured contaminant effects. It would be interesting to determine if the results of this study could be replicated in a more controlled setting (i.e. with known PCB exposures) as well as investigate if contaminant effects respond in a dose-related manner. Given the variation I saw in adult females, I would also expect in utero PCB exposure in this study to vary as well. It is unknown how maternal condition corresponds to the efficiency of PCB transfer (i.e. are mothers with higher lipid reserves able to offload PCBs at higher rates), which could then be a contributing factor to exposure embryo variability. Future studies should take care to measure contaminants in mothers and embryos to be able to link effects to actual dose, and to compare responses among littermates considering that I found evidence of resource competition in utero.

142

Figure 7.1. Elasmobranch PCB burdens for a range of species. Comparison of PCB concentrations measured in Round Stingrays from the current study to other elasmobranchs reported in the literature for both wet weight (A) and lipid weight (B) standardized concentrations. Round Stingrays are highlighted in dark grey relative to two other species (medium grey) that have been studied for contaminant effects, but whose results were inconclusive (Gelsleichter et al., 2006, 2005; Manire, 2002).

143

Histotrophy represents only one strategy among a range of matrotrophic modes utilized by elasmobranchs. The delivery system by which females transfer nutrients to their offspring could have consequences for intrauterine growth variability as well as contaminant exposure in utero. In histotrophy, maternal secretions of a nutritious fluid shared by all embryos represents a “free for all” situation within each uterus (left and right) where dominant embryos may be better able to outcompete their littermates. I would predict similar scenarios for elasmobranchs utilizing oophagy, where mothers, to our knowledge, indiscriminately ovulate unfertilized eggs into the uterus that are then consumed by embryos. In contrast, embryos of placental species typically develop in their own compartments, and each has their own, individual connection to their mother. This then represents a situation where mothers may have more control over the distribution of resources. It would be interesting to determine if intrauterine variability in mass is related to matrotrophic mode, with the prediction that strategies where mothers have more control, would lead to less competition and more equally-sized embryos. Mechanisms that facilitate lipid transfer (i.e. oophagy, lipid histotrophy) would be predicted to also ease the transfer of PCBs to embryos, thus creating a situation where the magnitude and rate of in utero exposure is dependent on the matrotrophic strategy of the species. Thus, differences in maternal provisioning method could have large implications for the extent of contaminant effects on embryos. One unforeseen challenge was capturing male stingrays as easily as pregnant females. This limited our ability to make specific sex comparisons each month of gestation to determine if disparities between males and females were exacerbated over the course of pregnancy. Despite this, our findings point to sex-specific responses to organochlorine contaminant exposure and the need to consider sex as a factor in further studies. Males appear to be more negatively impacted than females, which has implications for fitness differences between the sexes. Prior to birth, male embryos from the contaminated site, in particular, were less ‘fit’ than their reference counterparts. The effect of males being relatively more impaired by contaminant exposure also was found in adults, suggesting that contaminant exposure in utero has lifelong implications. Future work should investigate the point at which males and females are most vulnerable to sex- related contaminant effects. In particular, future work should focus on the point around maturity, as this is the time where male and female priorities diverge and sex-related differences in contaminant effects may become apparent.

144

While adult males from the contaminant-exposed site had clear metabolic consequences for males, I was unable to address other important questions regarding effects on reproductive fitness. Our males were sampled during the beginning of the recrudescent period (Mull et al., 2008), but not at the height of testicular growth and testosterone production. As such, I found no difference between males for steroid hormone concentrations, particularly for testosterone. However, to assess potential reproductive impairments sampling of males during their steroidogenic peak is needed. In addition, assessing testicular growth with respect to contaminant exposure is important as the energetic strains I documented in males could have cascading effects on the amount of energetic resources available for stingrays to invest in spermatogenesis. Considering that male competition appears to be relatively high in stingrays (Lyons et al., 2017), reduced testicular performance either directly or indirectly by contaminants could have significant effects on individual fitness. While Round Stingray populations are not decreasing, it begs the question if PCB exposure is having population-level implications. This is unlikely the case for the Round Stingray, however, contaminant exposure and effect should not be discounted in other, understudied elasmobranchs. An immediately identifiable negative effect of contaminant exposure in Round Stingrays was their inability to mount a robust response to an acute stressor. I found females at both sites mounting a primary stress response, but I was unable to confirm if sex played a role as I had too few samples of baseline males from the reference site. It would be interesting to repeat this experiment with a full complement of male and female experimental groups to quantify with statistical rigor the influence of sex on the primary stress response pathway. While our sampling period was constricted by the need to reduce stress effects on other parameters, future studies should further investigate the degree to which contaminant-exposure impairs elasmobranch stress physiology. During our ~15 min window, I found that reference stingrays were able to mobilize glucose from the liver into the plasma. Repeating this experiment with serial sampling of individuals over a longer period of time (i.e. 24 hr) is necessary to determine if the secondary response is completely abolished in contaminant-exposed stingrays or severely attenuated. In addition, further exploring changes in other energy substrates, particularly the role of ketones which was not well established in this study, is worthwhile as well considering that the metabolic organization of elasmobranchs differs from other vertebrates. The fitness effect (i.e. ability to escape predator, capacity to recover from fishing capture, etc.) of the difference between a

145 dampened and abolished response is unknown in elasmobranchs. This could play a critical role for species which experience extended capture times (i.e. commercial longlining) and also have a propensity to accumulate high contaminant burdens.

References Abdel-Aziz, S.H., El-Nady, F.S., 1993. Lipid dynamics in the common torpedo, Torpedo torpedo, from the south eastern Mediterranean. J. Fish Biol. 43, 155–162. Adams, S.M., Crumby, W.D., Greeley, M.S., Ryon, M.G., Schilling, E.M., 1992. Relationships between physiological and fish population responses in a contaminated stream. Environ. Toxicol. Chem. 11, 1549–1557. Adkins-Regan, E., Ottinger, M.A., Park, J., 1995. Maternal transfer of estradiol to egg yolks alters sexual differentiation of avian offspring. J. Exp. Zool. 271, 466–470. Allen, M., Diehl, D., Zeng, E., 2004. Bioaccumulation of contaminants in recreational and forage fish in Newport Bay, California in 2000–2002. Calif. Coast. Water Res. Aluru, N., Jorgensen, E.H., Maule, A.G., Vijayan, M.M., 2004. PCB disruption of the hypothalamus-pituitary-interrenal axis involves brain glucocorticoid receptor downregulation in anadromous Arctic charr. Am. J. Physiol. Regul. Integr. Comp. Physiol. 287, R787-R793. Aluru, N., Renaud, R., Leatherland, J.F., Vijayan, M.M., 2005. Ah receptor-mediated impairment of interrenal steroidogenesis involves StAR protein and P450scc gene attenuation in Rainbow Trout. Toxicol. Sci. 84, 260–269. Aluru, N., Vijayan, M.M., 2006. Aryl hydrocarbon receptor activation impairs cortisol response to stress in Rainbow Trout by disrupting the rate-limiting steps in steroidogenesis. Endocrinology 147, 1895–1903. Alves, L.M.F., Nunes, M., Marchand, P., Le Bizec, B., Mendes, S., Correia, J.P.S., Lemos, M.F.L., Novais, S.C., 2016. Blue sharks (Prionace glauca) as bioindicators of pollution and health in the Atlantic Ocean: contamination levels and biochemical stress responses. Sci. Total Environ. 563, 282–292. Amara, R., Méziane, T., Gilliers, C., Hermel, G., Laffargue, P., 2007. Growth and condition indices in juvenile sole Solea solea measured to assess the quality of essential fish habitat. Mar. Ecol. Prog. Ser. 351, 201–208. American Academy of Pediatrics Committee on Environmental Health., 2003. Polychlorinated biphenyls, dibenzofurans, and dibenzodioxins. Handb. Pediatr. Environ. Heal. 21, 215–22. Anderson, P.M., 2001. Urea and glutamine synthesis: Environmental influences on nitrogen excretion, in: Physiology, B.T.-F. (Ed.), Nitrogen Excretion. Academic Press, pp. 239–277. Anderson, P.M., 1990. Ketone body and phosphoenolpyruvate formation by isolated hepatic mitochondria from Squalus acanthias (spiny dogfish). J. Exp. Zool. Part A Ecol. Genet. Physiol. 254, 144–154. Anderson, P.M., 1986. Support of respiration and citrulline synthesis by isolated hepatic mitochondria from Squalus acanthias by acyl-CoAs and other nitrogen-donating substrates. Comp. Biochem. Physiol. Part B Comp. Biochem. 85, 789–794. Anderson, W.G., 2012. The endocrinology of 1α-hydroxycorticosterone in elasmobranch fish: A review. Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 162, 73–80. Anderson, W.G., Good, J.P., Pillans, R.D., Hazon, N., Franklin, C.E., 2005. Hepatic urea

146

biosynthesis in the euryhaline elasmobranch Carcharhinus leucas. J. Exp. Zool. Part A Ecol. Genet. Physiol. 303, 917–921. Armour, K.J., O’Toole, L.B., Hazon, N., 1993. The effect of dietary protein restriction on the secretory dynamics of 1α-hydroxycorticosterone and urea in the dogfish, Scyliorhinus canicula: a possible role for 1α-hydroxycorticosterone in sodium retention. J. Endocrinol. 138, 275–282. Axelrod, J., Reisine, T.D., 1984. Stress hormones- their interaction and regulation. Science 224, 452–459. Aydin, Y., Erkan, M., 2017. The toxic effects of polychlorinated biphenyl (Aroclor 1242) on Tm3 Leydig cells. Toxicol. Ind. Health 748233717699783. Babel, J.S., 1967. Reproduction, life history, and ecology of the round stingray, Urolophus halleri Cooper. Fish Bull. 1–104. Baker, L.N., Chapman, A.B., Grummer, R.H., Casida, L.E., 1958. Some factors affecting litter size and fetal weight in purebred and reciprocal-cross matings of Chester White and Poland China swine. J. Anim. Sci. 17, 612–621. Ballantyne, J.S., 1997. Jaws: The Inside Story. The Metabolism of Elasmobranch Fishes. Comp. Biochem. Physiol. Part B Biochem. Mol. Biol. 118, 703–742. Barcellos, L.J.G., Nicolaiewsky, S., De Souza, S.M.G., Lulhier, F., 1999. Plasmatic levels of cortisol in the response to acute stress in Nile tilapia, Oreochromis niloticus (L.), previously exposed to chronic stress. Aquac. Res. 30, 437–444. Barton, B.A., 2002. Stress in fishes: A diversity of responses with particular reference to changes in circulating corticosteroids. Integr. Comp. Biol. 42, 517–525. Bazer, F.W., Roberts, R.M., Thatcher, W.W., 1979. Actions of hormones on the uterus and effect on conceptus development. J. Anim. Sci. 49, 35–45. Bergeron, J.M., Crews, D., McLachlan, J.A., 1994. PCBs as environmental estrogens: turtle sex determination as a biomarker of environmental contamination. Environ. Health Perspect. 102, 780. Bergmeyer, H.U., 1983. No Title, 3rd ed, Methods of Enzymatic Analysis. Beyers, D.W., Rice, J.A., Clements, W.H., Henry, C.J., 1999. Estimating physiological cost of chemical exposure: integrating energetics and stress to quantify toxic effects in fish. Can. J. Fish. Aquat. Sci. 56, 814–822. Binnington, M.J., Wania, F., 2014. Clarifying relationships between persistent organic pollutant concentrations and age in wildlife biomonitoring: individuals, cross-sections, and the roles of lifespan and sex. Environ. Toxicol. Chem. 33, 1415–1426. Birnbaum, L.S., Tuomisto, J., 2000. Non-carcinogenic effects of TCDD in animals. Food Addit. Contam. 17, 275–288. Black, D.E., Phelps, D.K., Lapan, R.L., 1988. The effect of inherited contamination on egg and larval winter flounder, Pseudopleuronectes americanus. Mar. Environ. Res. 25, 45–62. Blasius, M.E., Goodmanlowe, G.D., 2008. Contaminants still high in top-level carnivores in the Southern California Bight: Levels of DDT and PCBs in resident and transient pinnipeds. Mar. Pollut. Bull. 56, 1973–1982. Bonga, S., 1997. The stress response in fish. Physiol. Rev. 77, 591-625. Borgå, K., Kidd, K.A., Muir, D.C.G., Berglund, O., Conder, J.M., Gobas, F.A.P.C., Kucklick, J., Malm, O., Powell, D.E., 2012. Trophic magnification factors: Considerations of ecology, ecosystems, and study design. Integr. Environ. Assess. Manag. 8, 64–84. Borrell, A., Bloch, D., Desportes, G., 1995. Age trends and reproductive transfer of

147

organochlorine compounds in long-finned pilot whales from the Faroe Islands. Environ. Pollut. 88, 283–292. Bowden, R.M., Ewert, M.A., Lipar, J.L., Nelson, C.E., 2001. Concentrations of steroid hormones in layers and biopsies of chelonian egg yolks. Gen. Comp. Endocrinol. 121, 95–103. Brar, N.K., Waggoner, C., Reyes, J.A., Fairey, R., Kelley, K.M., 2010. Evidence for thyroid endocrine disruption in wild fish in San Francisco Bay, California, USA. Relationships to contaminant exposures. Aquat. Toxicol. 96, 203–215. Breuner, C.W., Patterson, S.H., Hahn, T.P., 2008. In search of relationships between the acute adrenocortical response and fitness. Gen. Comp. Endocrinol. 157, 288–295. Brown, S.B., Adams, B.A., Cyr, D.G., Eales, J.G., 2004. Contaminant effects on the teleost fish thyroid. Environ. Toxicol. Chem. 23, 1680–1701. Bugel, S.M., White, L.A., Cooper, K.R., 2011. Decreased vitellogenin inducibility and 17β- estradiol levels correlated with reduced egg production in killifish (Fundulus heteroclitus) from Newark Bay, NJ. Aquat. Toxicol. 105, 1–12. Bursian, S.J., Kern, J., Remington, R.E., Link, J.E., Fitzgerald, S.D., 2013. Dietary exposure of mink (Mustela vison) to fish from the upper Hudson River, New York, USA: Effects on reproduction and offspring growth and mortality. Environ. Toxicol. Chem. 32, 780–793. Camhi, M.D., Valenti, S. V, Fordham, S. V, Fowler, S.L., Gibson, C., 2009. The conservation status of pelagic sharks and rays: report of the IUCN shark specialist group pelagic shark red list workshop. IUCN Species Surviv. Comm. Shark Spec. Group. Newbury, UK. x+ 78p. Castro, J., 1983. The Sharks of North American Waters. Texas A&AM University Press, College Station, TX. Chandra, R., Chaudhary, S., 2013. Persistent organic pollutants in environment and their health hazards. Int. J. Bioassays 2, 1232–1238. Clarke, S. (2013). Towards an integrated shark conservation and management measure for the Western and Central Pacific Ocean. WCPFC-SC9-2013/EB-WP-08. WCPFC-SC, Pohnpei, Federated States of Micronesia, 6-14 August. Close, D.A., Yun, S.-S., McCormick, S.D., Wildbill, A.J., Li, W., 2010. 11-Deoxycortisol is a corticosteroid hormone in the lamprey. Proc. Natl. Acad. Sci. 107, 13942–13947. Cockrem, J.F., 2013. Individual variation in glucocorticoid stress responses in animals. Gen. Comp. Endocrinol. 181, 45–58. Compagno, L.J. V, 1990. Alternative life-history styles of cartilaginous fishes in time and space, in: Alternative Life-History Styles of Fishes. Springer, pp. 33–75. Conley, A.J., Elf, P., Corbin, C.J., Dubowsky, S., Fivizzani, A., Lang, J.W., 1997. Yolk steroids decline during sexual differentiation in the alligator. Gen. Comp. Endocrinol. 107, 191–200. Crain, D.A., Guillette, L.J., Pickford, D.B., Percival, H.F., Woodward, A.R., 1998. Sex-steroid and thyroid hormone concentrations in juvenile alligators (Alligator mississippiensis) from contaminated and reference lakes in Florida, USA. Environ. Toxicol. Chem. 17, 446–452. Crear, D., 2015. The effects of temperature and habitat on the movement patterns of green sea turtles (Chelonia mydas) within an urbanized river and local estuary. California State University, Long Beach. Crespi, B., Semeniuk, C., 2004. Parent-offspring conflict in the evolution of vertebrate reproductive mode. Am. Nat. 163, 635–653. Crisp, T.M., Clegg, E.D., Cooper, R.L., Wood, W.P., Anderson, D.G., Baetcke, K.P., Hoffmann, J.L., Morrow, M.S., Rodier, D.J., Schaeffer, J.E., Touart, L.W., Zeeman, M.G., Patel, Y.M.,

148

1998. Environmental endocrine disruption: an effects assessment and analysis. Environ. Health Perspect. 106, 11–56. Davis, K.B., Suttle, M.A., Parker, N.C., 1984. Biotic and abiotic influences on corticosteroid hormone rhythms in channel catfish. Trans. Am. Fish. Soc. 113, 414–421. De Boeck, G., Hattink, J., Franklin, N.M., Bucking, C.P., Wood, S., Walsh, P.J., Wood, C.M., 2007. Copper toxicity in the spiny dogfish (Squalus acanthias): Urea loss contributes to the osmoregulatory disturbance. Aquat. Toxicol. 84, 133–141. de Souza Valentim, F.C., Porto, J.I.R., Bertollo, L.A.C., Gross, M.C., Feldberg, E., 2013. XX/XO, a rare sex chromosome system in Potamotrygon freshwater stingray from the Amazon Basin, Brazil. Genetica 141, 381–387. De Tata, V., 2014. Association of dioxin and other persistent organic pollutants (POPs) with diabetes: Epidemiological evidence and new mechanisms of beta cell dysfunction. Int. J. Mol. Sci. 15, 7787–7811. Debier, C., Pomeroy, P.P., Dupont, C., Joiris, C., Comblin, V., Le Boulengé, E., Larondelle, Y., Thomé, J.P., 2003. Quantitative dynamics of PCB transfer from mother to pup during lactation in UK grey seals Halichoerus grypus. Mar. Ecol. Prog. Ser. 247, 237–248. Duntas, L.H., 2015. Chemical contamination and the thyroid. Endocrine 48, 53–64. Eganhouse, R.P., Pontolillo, J., 2000. Depositional history of organic contaminants on the Palos Verdes Shelf, California. Mar. Chem. 70, 317–338. Eganhouse, R.P., Pontolillo, J., Leiker, T.J., 2000. Diagenetic fate of organic contaminants on the Palos Verdes Shelf, California. Mar. Chem. 70, 289–315. Eisenreich, K.M., Kelly, S.M., Rowe, C.L., 2009. Latent mortality of juvenile snapping turtles from the upper Hudson River, New York, exposed maternally and via the diet to polychlorinated biphenyls (PCBs). Environ. Sci. Technol. 43, 6052–6057. Eisler, R., Belisle, A.A., 1996. Planar PCB hazards to fish, wildlife, and invertebrates: A synoptic review. Patuxent Wildlife Research Center, Laurel, MD, report no. 31. Elgar, M.A., 1990. Evolutionary compromise between a few large and many small eggs: comparative evidence in teleost fish. Oikos 59, 283–287. Elonen, G.E., Spehar, R.L., Holcombe, G.W., Johnson, R.D., Fernandez, J.D., Erickson, R.J., Tietge, J.E., Cook, P.M., 1998. Comparative toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin to seven freshwater fish species during early life-stage development. Environ. Toxicol. Chem. 17, 472–483. Evans, A.N., Rimoldi, J.M., Gadepalli, R.S. V, Nunez, B.S., 2010. Adaptation of a corticosterone ELISA to demonstrate sequence-specific effects of angiotensin II peptides and C-type natriuretic peptide on 1α-hydroxycorticosterone synthesis and steroidogenic mRNAs in the elasmobranch interrenal gland. J. Steroid Biochem. Mol. Biol. 120, 149–154. Evans, D.H., 1987. The fish gill: site of action and model for toxic effects of environmental pollutants. Environ. Health Perspect. 71, 47. Evans, D.H., Piermarini, P.M., Choe, K.P., 2004. Homeostasis: osmoregulation, pH regulation, and nitrogen excretion, in: Carrier, J.C., Musick, J.A., Heithaus, M.R. (Eds.), Biology of Sharks and Their Relatives. CRC Press, Boca Raton, pp. 247–268. Fasano, S., D’Antonio, M., Pierantoni, R., Chieffi, G., 1992. Plasma and follicular tissue steroid levels in the elasmobranch fish, Torpedo marmorata. Gen. Comp. Endocrinol. 85, 327–333. Feist, G.W., Webb, M.A.H., Gundersen, D.T., Foster, E.P., Schreck, C.B., Maule, A.G., Fitzpatrick, M.S., 2005. Evidence of detrimental effects of environmental contaminants on growth and reproductive physiology of white sturgeon in impounded areas of the Columbia

149

River. Environ. Health Perspect. 113, 1675–1682. Fernie, K., Bortolotti, G., Drouillard, K., Smits, J., Marchant, T., 2003a. Developmental toxicity of in ovo exposure to polychlorinated biphenyls: II. Effects of maternal or paternal exposure on second-generation nestling American. Environ. Toxicol. Chem. 22, 2688–2694. Fernie, K., Smits, J., Bortolotti, G., 2003b. Developmental toxicity of in ovo exposure to polychlorinated biphenyls: I. Immediate and subsequent effects on first-generation nestling American kestrels (Falco sparverius). Environ. Toxicol. Chem. 22, 554–560. Fiandanese, N., Borromeo, V., Berrini, A., Fischer, B., Schaedlich, K., Schmidt, J.-S., Secchi, C., Pocar, P., 2016. Maternal exposure to a mixture of di(2-ethylhexyl) phthalate (DEHP) and polychlorinated biphenyls (PCBs) causes reproductive dysfunction in adult male mouse offspring. Reprod. Toxicol. 65, 123–132. Fines, G.A., Ballantyne, J.S., Wright, P.A., 2001. Active urea transport and an unusual basolateral membrane composition in the gills of a marine elasmobranch. Am. J. Physiol. - Regul. Integr. Comp. Physiol. 280, R16--R24. Fisk, A.T., Tittlemier, S. a, Pranschke, J.L., Norstrom, R.J., 2002. Using anthropogenic contaminants and stable isotopes to assess the feeding ecology of Greenland Sharks. Ecology 83, 2162–2172. Folch, J., Lees, M., Sloane-Stanley, G.H., others, 1957. A simple method for the isolation and purification of total lipids from animal tissues. J Biol Chem 226, 497–509. Folland, W.R., Newsted, J.L., Fitzgerald, S.D., Fuchsman, P.C., Bradley, P.W., Kern, J., Kannan, K., Remington, R.E., Zwiernik, M.J., 2016. Growth and reproductive effects from dietary exposure to Aroclor 1268 in mink (Neovison vison), a surrogate model for marine mammals. Environ. Toxicol. Chem. 35, 604–618. doi:10.1002/etc.3201 Forde, N., Carter, F., Fair, T., Crowe, M.A., Evans, A.C.O., Spencer, T.E., Bazer, F.W., McBride, R., Boland, M.P., O’Gaora, P., Lonergan, P., Roche, J.F., 2009. Progesterone- regulated changes in endometrial gene expression contribute to advanced conceptus development in cattle. Biol. Reprod. 81, 784–794. doi:10.1095/biolreprod.108.074336 Forster, R.P., Goldstein, L., 1969. Formation of excretory products, in: Hoar, W., Randall, D. (Eds.), Fish Physiology. Academic Press, New York, pp. 313–350. Frantz, D.E., 2014. The effects of Aroclor 1254 on the reproductive function of male round stingrays (Urobatis halleri). California State University Long Beach. Frisk, M.G., Miller, T.J., Fogarty, M.J., 2001. Estimation and analysis of biological parameters in elasmobranch fishes: a comparative life history study. Can. J. Fish. Aquat. Sci. 58, 969– 981. doi:10.1139/f01-051 Gelsleichter, J., Manire, C.A., Szabo, N.J., Cortés, E., Carlson, J., Lombardi-Carlson, L., 2005. Organochlorine concentrations in bonnethead sharks (Sphyrna tiburo) from four Florida estuaries. Arch. Environ. Contam. Toxicol. 48, 474–483. Gelsleichter, J., Walsh, C.J., Szabo, N.J., Rasmussen, L.E.L., 2006. Organochlorine concentrations, reproductive physiology, and immune function in unique populations of freshwater Atlantic stingrays (Dasyatis sabina) from Florida’s St. Johns River. Chemosphere 63, 1506–1522. Glickman, S.E., Frank, L.G., Davidson, J.M., Smith, E.R., Siiteri, P.K., 1987. Androstenedione may organize or activate sex-reversed traits in female spotted hyenas. Proc. Natl. Acad. Sci. 84, 3444–3447. Gobas, F.A.P.C., Arnot, J.A., 2010. Food web bioaccumulation model for polychlorinated biphenyls in San Francisco Bay, California, USA. Environ. Toxicol. Chem. 29, 1385–1395.

150

Goldey, E.S., Kehn, L.S., Lau, C., Rehnberg, G.L., Crofton, K.M., 1995. Developmental exposure to polychlorinated biphenyls (Aroclor 1254) reduces circulating thyroid hormone concentrations and causes hearing deficits in rats. Toxicol. Appl. Pharmacol. 135, 77–88. Goldstein, L., Forster, R.P., 1971. Osmoregulation and urea metabolism in the little skate Raja erinacea. Am. J. Physiol. 220, 742–746. Goldstein, L., Funkhouser, D., 1972. Biosynthesis of trimethylamine oxide in the nurse shark, Ginglymostoma cirratum. Comp. Biochem. Physiol. Part A Physiol. 42, 51–57. doi:https://doi.org/10.1016/0300-9629(72)90365-9 Goldstein, L., Hartman, S.C., Forster, R.P., 1967. On the origin of trimethylamine oxide in the spiny dogfish, Squalus acanthias. Comp. Biochem. Physiol. 21, 719–722. Guillette, L.J., Gross, T.S., Masson, G.R., Matter, J.M., Percival, H.F., Woodward, A.R., 1994. Developmental abnormalities of the gonad and abnormal sex hormone concentrations in juvenile alligators from contaminated and control lakes in Florida. Environ. Health Perspect. 102, 680–688. Hale, L.F., Dudgeon, J. V., Mason, A.Z., Lowe, C.G., 2006. Elemental signatures in the vertebral cartilage of the round stingray, Urobatis halleri, from Seal Beach, California. Environ. Biol. Fishes 77, 317–325. doi:10.1007/s10641-006-9124-2 Hamlett, W.C., Eulitt, A.M., Jarrell, R.L., Kelly, M.A., 1993. Uterogestation and placentation in elasmobranchs. J. Exp. Zool. Part A Ecol. Genet. Physiol. 266, 347–367. Hany, J., Lilienthal, H., Sarasin, A., Roth-Härer, A., Fastabend, A., Dunemann, L., Lichtensteiger, W., Winneke, G., 1999. Developmental exposure of rats to a reconstituted PCB mixture or Aroclor 1254: effects on organ weights, aromatase activity, sex hormone levels, and sweet preference behavior. Toxicol. Appl. Pharmacol. 158, 231–243. Heithaus, M.R., Frid, A., Wirsing, A.J., Worm, B., 2008. Predicting ecological consequences of marine top predator declines. Trends Ecol. Evol. 23, 202–210. Hertz-Picciotto, I., Park, H.-Y., Dostal, M., Kocan, A., Trnovec, T., Sram, R., 2008. Prenatal exposures to persistent and non-persistent organic compounds and effects on immune system development. Basic Clin. Pharmacol. Toxicol. 102, 146–154. Hickie, B.E., Ross, P.S., Macdonald, R.W., Ford, J.K., 2007. Killer whales (Orcinus orca) face protracted health risks associated with lifetime exposure to PCBs. Environ. Sci. Technol. 41, 6613–9. Hight, B. V, Lowe, C.G., 2007. Elevated body temperatures of adult female leopard sharks, Triakis semifasciata, while aggregating in shallow nearshore embayments: Evidence for behavioral thermoregulation? J. Exp. Mar. Bio. Ecol. 352, 114–128. Hoff, G.R., 2009. Embryo developmental events and the egg case of the Aleutian skate Bathyraja aleutica (Gilbert) and the Alaska skate Bathyraja parmifera (Bean). J. Fish Biol. 74, 483–501. Hoffman, D.J., Rattner, B.A., Bunck, C.M., Krynitsky, A., Ohlendorf, H.M., Lowe, R.W., 1986. Association between PCBs and lower embryonic weight in black-crowned night herons in San Francisco bay. J. Toxicol. Environ. Health 19, 383–391. Hoffmayer, E.R., Hendon, J.M., Parsons, G.R., 2012. Seasonal modulation in the secondary stress response of a carcharhinid shark, Rhizoprionodon terraenovae. Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 162, 81–87. Hoisington, G., Lowe, C.G., 2005. Abundance and distribution of the round stingray, Urobatis halleri, near a heated effluent outfall. Mar. Environ. Res. 60, 437–453. Hontela, A., Dumont, P., Duclos, D., Fortin, Réj., 1995. Endocrine and metabolic dysfunction in

151

yellow perch, Perca flavescens, exposed to organic contaminants and heavy metals in the ST. Lawrence river. Environ. Toxicol. Chem. 14, 725–731. Hontela, A., Rasmussen, J.B., Audet, C., Chevalier, G., 1992. Impaired cortisol stress response in fish from environments polluted by PAHs, PCBs, and mercury. Arch. Environ. Contam. Toxicol. 22, 278–283. Hontela, A., Vijayan, M.M., 2008. Adrenocortical toxicology in fishes, in: Harvey, P., Everett, D., Springall, C. (Eds.), Adrenal Toxicology. Taylor and Francis, Boca Raton, Florida, pp. 233–256. Hose, J.E., Cross, J.N., Smith, S.G., Diehl, D., 1989. Reproductive impairment in a fish inhabiting a contaminated coastal environment off Southern California. Environ. Pollut. 57, 139–148. Hussey, N.E., Wintner, S.P., Dudley, S.F.J., Cliff, G., Cocks, D.T., Aaron MacNeil, M., 2010. Maternal investment and size-specific reproductive output in carcharhinid sharks. J. Anim. Ecol. 79, 184–193. Idler, D.R., Truscott, B., 1966. 1α-hydroxycorticosterone from cartilaginous fish: a new adrenal steroid in blood. J. Fish. Board Canada 23, 615–619. Ings, J.S., Servos, M.R., Vijayan, M.M., 2011. Hepatic transcriptomics and protein expression in Rainbow Trout exposed to municipal wastewater effluent. Environ. Sci. Technol. 45, 2368– 2376. Jansen, H.T., Cooke, P.S., Porcelli, J., Liu, T.-C., Hansen, L.G., 1993. Estrogenic and antiestrogenic actions of PCBs in the female rat: in vitro and in vivo studies. Reprod. Toxicol. 7, 237–248. Janzen, F.J., Wilson, M.E., Tucker, J.K., Ford, S.P., 1998. Endogenous yolk steroid hormones in turtles with different sex-determining mechanisms. Gen. Comp. Endocrinol. 111, 306–317. Jirik, K.E., Lowe, C.G., 2012. An elasmobranch maternity ward: female round stingrays Urobatis halleri use warm, restored estuarine habitat during gestation. J. Fish Biol. 80, 1227–1245. Johnson, L.L., Casillas, E., Collier, T.K., McCain, B.B., Varanasi, U., 1988. Contaminant effects on ovarian development in English Sole (Parophrys vetulus) from Puget Sound, Washington. Can. J. Fish. Aquat. Sci. 45, 2133–2146. doi:10.1139/f88-248 Jonsson, N., Jonsson, B., Hansen, L.P., 1997. Changes in proximate composition and estimates of energetic costs during upstream migration and spawning in Atlantic salmon Salmo salar. J. Anim. Ecol. 425–436. Kaya, H., Hany, J., Fastabend, A., Roth-Härer, A., Winneke, G., Lilienthal, H., 2002. Effects of maternal exposure to a reconstituted mixture of polychlorinated biphenyls on sex-dependent behaviors and steroid hormone concentrations in rats: dose–response relationship. Toxicol. Appl. Pharmacol. 178, 71–81. King, K.J., 1984. Changes in condition of mature female rig (Mustelus lenticulatus) from Golden Bay in relation to seasonal inshore migrations. New Zeal. J. Mar. Freshw. Res. 18, 21–27. Klasing, S., Witting, D., Brodberg, R., Gassel, M., 2009. Health advisory and safe eating guidelines for fish from coastal areas of southern California: Ventura Harbor to San Mateo Point. Off. Environ. Heal. Hazard Assessment, Sacramento, CA 42. Kormanik, G.A., 1993. Ionic and osmotic environment of developing elasmobranch embryos, in: The Reproduction and Development of Sharks, Skates, Rays and Ratfishes. Springer, pp. 233–240. Kormanik, G.A., 1992. Ion and osmoregulation in prenatal elasmobranchs: evolutionary

152

implications. Am. Zool. 32, 294–302. Larsson, Å., Fänge, R., 1977. Cholesterol and free fatty acids (FFA) in the blood of marine fish. Comp. Biochem. Physiol. Part B Comp. Biochem. 57, 191–196. Lauby-Secretan, B., Loomis, D., Baan, R., El Ghissassi, F., Bouvard, V., Benbrahim-Tallaa, L., Guha, N., Grosse, Y., Straif, K., 2016. Use of mechanistic data in the IARC evaluations of the carcinogenicity of polychlorinated biphenyls and related compounds. Environ. Sci. Pollut. Res. 23, 2220–2229. doi:10.1007/s11356-015-4829-4 Lee, J.S., Tanabe, S., Umino, H., Tatsukawa, R., Loughlin, T.R., Calkins, D.C., 1996. Persistent organochlorines in Steller sea lion (Eumetopias jubatus) from the bulk of Alaska and the Bering Sea, 1976--1981. Mar. Pollut. Bull. 32, 535–544. Leijs, M.M., van der Linden, L.M., Koppe, J.G., Olie, K., van Aalderen, W.M.C., ten Tusscher, G.W., 2014. The influence of perinatal and current dioxin and PCB exposure on reproductive parameters (sex-ratio, menstrual cycle characteristics, endometriosis, semen quality, and prematurity): a review. Biomonitoring 1, 1–15. Levesque, H.M., Moon, T.W., Campbell, P.G.C., Hontela, A., 2002. Seasonal variation in carbohydrate and lipid metabolism of yellow perch (Perca flavescens) chronically exposed to metals in the field. Aquat. Toxicol. 60, 257–267. Love, O.P., Shutt, L.J., Silfies, J.S., Bortolotti, G.R., Smits, J.E.G., Bird, D.M., 2003. Effects of Dietary PCB Exposure on Adrenocortical Function in Captive American Kestrels (Falco sparverius). Ecotoxicology 12, 199–208. Lowe, C.G., Moss, G.J., Hoisington, G., Vaudo, J.J., Cartamil, D.P., Marcotte, M.M., Papastamatiou, Y.P., 2007. Caudal Spine Shedding Periodicity and Site Fidelity of Round Stingrays, Urobatis halleri (Cooper), at Seal Beach, California: Implications for Stingray- related Injury Management. Bull. South. Calif. Acad. Sci. 106, 16–26. Lundin, J.I., Ylitalo, G.M., Booth, R.K., Anulacion, B., Hempelmann, J.A., Parsons, K.M., Giles, D.A., Seely, E.A., Hanson, M.B., Emmons, C.K., others, 2016. Modulation in persistent organic pollutant concentration and profile by prey availability and reproductive status in Southern Resident killer whale scat samples. Environ. Sci. Technol. 50, 6506–6516. Lyons, K., Adams, D.H., 2015. Maternal offloading of organochlorine contaminants in the yolk- sac placental scalloped hammerhead shark (Sphyrna lewini). Ecotoxicology 24, 553–562. Lyons, K., Carlisle, A., Preti, A., Mull, C., Blasius, M., O’Sullivan, J., Winkler, C., Lowe, C.G., 2013. Effects of trophic ecology and habitat use on maternal transfer of contaminants in four species of young of the year lamniform sharks. Mar. Environ. Res. 90, 27–38. Lyons, K., Chabot, C.L., Mull, C.G., Paterson Holder, C.N., Lowe, C.G., 2017. Who’s My Daddy? Considerations for the influence of sexual selection on multiple paternity in elasmobranch mating systems. Ecol. Evol. 1–10. Lyons, K., Lavado, R., Schlenk, D., Lowe, C.G., 2014. Bioaccumulation of organochlorine contaminants and ethoxyresorufin-o-deethylase activity in southern California round stingrays (Urobatis halleri) exposed to planar aromatic compounds. Environ. Toxicol. Chem. 33, 1380–1390. doi:10.1002/etc.2564 Lyons, K., Lowe, C.G., 2015. Organochlorine contaminants and maternal offloading in the lecithotrophic Pacific angel shark (Squatina californica) collected from southern California. Mar. Pollut. Bull. 97, 518–522. Lyons, K., Lowe, C.G., 2013a. Quantification of maternal offloading of organic contaminants in elasmobranchs using the histotrophic round stingray (Urobatis halleri) as a model. Environ. Sci. Technol. 47, 12450–12458.

153

Lyons, K., Lowe, C.G., 2013b. Mechanisms of maternal transfer of organochlorine contaminants and mercury in the common thresher shark (Alopias vulpinus). Can. J. Fish. Aquat. Sci. 70, 1667–1672. Maddock, M.B., Schwartz, F.J., 1996. Elasmobranch cytogenetics: methods and sex chromosomes. Bull. Mar. Sci. 58, 147–155. Madenjian, C.P., Rediske, R.R., Krabbenhoft, D.P., Stapanian, M.A., Chernyak, S.M., O’Keefe, J.P., 2016. Sex differences in contaminant concentrations of fish: a synthesis. Biol. Sex Differ. 7, 42. Manire, C.A., 2002. The mechanisms and effects of endocrine disruption on infertility in the Bonnethead Shark on Florida’s Gulf Coast. Mote Marine Laboratory, Sarasota, FL. Manire, C.A., Rasmussen, L.E.L., Gelsleichter, J., Hess, D.L., 2004. Maternal serum and yolk hormone concentrations in the placental viviparous bonnethead shark, Sphyrna tiburo. Gen. Comp. Endocrinol. 136, 241–247. Manire, C.A., Rasmussen, L.E.L., Hess, D.L., Hueter, R.E., 1995. Serum steroid hormones and the reproductive cycle of the female Bonnethead Shark, Sphyrna tiburo. Gen. Comp. Endocrinol. 97, 366–376. Manire, C.A., Rasmussen, L.E.L., Maruska, K.P., Tricas, T.C., 2007. Sex, seasonal, and stress- related variations in elasmobranch corticosterone concentrations. Comp. Biochem. Physiol. - A Mol. Integr. Physiol. 148, 926–935. Manire, C., Hueter, R., Hull, E., Spieler, R., 2001. Serological changes associated with gill-net capture and restraint in three species of sharks. Trans. Am. Fish. Soc. 130, 1038–1048. Mann, G.E., Lamming, G.E., 1999. The influence of progesterone during early pregnancy in cattle. Reprod. Domest. Anim. 34, 269–274. Mappes, T., Koskela, E., Lopez-Fanjul, C., 2004. Genetic basis of the trade-off between offspring number and quality in the bank vole. Evolution 58, 645–650. Marchand, J., Quiniou, L., Riso, R., Thebaut, M.-T., Laroche, J., 2004. Physiological cost of tolerance to toxicants in the European flounder Platichthys flesus, along the French Atlantic Coast. Aquat. Toxicol. 70, 327–343. Marentette, J.R., Tong, S., Balshine, S., 2013. The cortisol stress response in male round goby (Neogobius melanostomus): effects of living in polluted environments? Environ. Biol. Fishes 96, 723–733. Marshall, H., Field, L., Afiadata, A., Sepulveda, C., Skomal, G., Bernal, D., 2012. Hematological indicators of stress in longline-captured sharks. Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 162, 121–129. Martin, A.P., Naylor, G.J.P., Palumbi, S.R., 1992. Rates of mitochondrial DNA evolution in sharks are slow compared with mammals. Nature 357, 153–155. Martyniuk, C.J., Doperalski, N.J., Prucha, M.S., Zhang, J.-L., Kroll, K.J., Conrow, R., Barber, D.S., Denslow, N.D., 2016. High contaminant loads in Lake Apopka’s riparian wetland disrupt gene networks involved in reproduction and immune function in largemouth bass. Comp. Biochem. Physiol. Part D Genomics Proteomics 19, 140–150. Matsumura, F., 1995. Mechanism of action of dioxin-type chemicals, pesticides, and other xenobiotics affecting nutritional indexes. Am. J. Clin. Nutr. 61, 695S--701S. Mauck, W.L., Mehrle, P.M., Mayer, F.L., 1978. Effects of the polychlorinated biphenyl Aroclor® 1254 on growth, survival, and bone development in brook trout (Salvelinus fontinalis). J. Fish. Board Canada 35, 1084–1088. McComb, D.M., Gelsleichter, J., Manire, C.A., Brinn, R., Brown, C.L., 2005. Comparative

154

thyroid hormone concentration in maternal serum and yolk of the bonnethead shark (Sphyrna tiburo) from two sites along the coast of Florida. Gen. Comp. Endocrinol. 144, 167–173. McCormick, S.D., 1993. Methods for Nonlethal Gill Biopsy and Measurement of Na+, K+- ATPase Activity. Can. J. Fish. Aquat. Sci. 50, 656–658. Merkel, K., Henderson, H., 2014. 2013 Anaheim Bay benthic and fish community assessments at the naval weapons station Seal Beach. Merkel and Associates. p 184. Mommsen, T.P., Vijayan, M.M., Moon, T.W., 1999. Cortisol in teleosts: dynamics, mechanisms of action, and metabolic regulation. Rev. Fish Biol. Fish. 9, 211–268. Mommsen, T.P., Walsh, P.J., 1991. Metabolic and enzymatic heterogeneity in the liver of the ureogenic teleost Opsanus beta. J. Exp. Biol. 156, 407–418. Moon, T.W., Mommsen, T.P., 1987. Enzymes of intermediary metabolism in tissue of the little skate, Raja erinacea. J. Exp. Zool. Part A Ecol. Genet. Physiol. 244, 9–15. Moore, I.T., Jessop, T.S., 2003. Stress, reproduction, and adrenocortical modulation in amphibians and reptiles. Horm. Behav. 43, 39–47. Moore, M.C., Johnston, G.I.H., 2008. Toward a dynamic model of deposition and utilization of yolk steroids. Integr. Comp. Biol. 48, 411–418. Morgan, R.L., Wright, P.A., Ballantyne, J.S., 2003. Urea transport in kidney brush-border membrane vesicles from an elasmobranch, Raja erinacea. J. Exp. Biol. 206, 3293–3302. Mull, C.G., Lowe, C.G., Young, K.A., 2010. Seasonal reproduction of female round stingrays (Urobatis halleri): Steroid hormone profiles and assessing reproductive state. Gen. Comp. Endocrinol. 166, 379–387. Mull, C.G., Lowe, C.G., Young, K.A., 2008. Photoperiod and water temperature regulation of seasonal reproduction in male round stingrays (Urobatis halleri). Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 151, 717–725. Mull, C.G., Lyons, K., Blasius, M.E., Winkler, C., O’Sullivan, J.B., Lowe, C.G., 2013. Evidence of maternal offloading of organic contaminants in White Sharks (Carcharodon carcharias). PLoS One 8, e62886. Murphy, V.E., Smith, R., Giles, W.B., Clifton, V.L., 2006. Endocrine Regulation of Human Fetal Growth: The Role of the Mother, Placenta, and Fetus. Endocr. Rev. 27, 141–169. Murugesan, P., Kanagaraj, P., Yuvaraj, S., Balasubramanian, K., Aruldhas, M.M., Arunakaran, J., 2005. The inhibitory effects of polychlorinated biphenyl Aroclor 1254 on Leydig cell LH receptors, steroidogenic enzymes and antioxidant enzymes in adult rats. Reprod. Toxicol. 20, 117–126. Musick, J.A., 1999. Life in the slow lane: ecology and conservation of long-lived marine animals. American Fisheries Society, Bethesda, Maryland, USA, Sewickly, PA, p. 265. Nagamani, M., McDonough, P.G., Ellegood, J.O., Mahesh, V.B., 1979. Maternal and amniotic fluid steroids throughout human pregnancy. Am. J. Obstet. Gynecol. 134, 674–680. Nebert, D.W., Roe, A.L., Dieter, M.Z., Solis, W.A., Yang, Y., Dalton, T.P., 2000. Role of the aromatic hydrocarbon receptor and [Ah] gene battery in the oxidative stress response, cell cycle control, and apoptosis. Biochem. Pharmacol. 59, 65–85. Niimi, A.J., 1983. Biological and toxicological effects of environmental contaminants in fish and their eggs. Can. J. Fish. Aquat. Sci. 40, 306–312. doi:10.1139/f83-045 O’Shaughnessy, K.L., Dahn, R.D., Cohn, M.J., 2015. Molecular development of chondrichthyan claspers and the evolution of copulatory organs. Nat. Commun. 6. Örn, S., Andersson, P.L., Förlin, L., Tysklind, M., Norrgren, L., 1998. The impact on

155

reproduction of an orally administered mixture of selected PCBs in zebrafish (Danio rerio). Arch. Environ. Contam. Toxicol. 35, 52–57. Pillans, R.D., Good, J.P., Gary Anderson, W., Hazon, N., Franklin, C.E., 2005. Freshwater to seawater acclimation of juvenile bull sharks (Carcharhinus leucas): Plasma osmolytes and Na+/K+-ATPase activity in gill, rectal gland, kidney and intestine. J. Comp. Physiol. B Biochem. Syst. Environ. Physiol. 175, 37–44. Plank, S.M., Lowe, C.G., Feldheim, K.A., Wilson, R.R., Brusslan, J.A., 2010. Population genetic structure of the round stingray Urobatis halleri (Elasmobranchii: Rajiformes) in southern California and the Gulf of California. J. Fish Biol. 77, 329–340. Pondella, D., Allen, L., 2008. The decline and recovery of four predatory fishes from the Southern California Bight. Mar. Biol. 154, 307–313. Quabius, E.S., Balm, P.H., Wendelaar Bonga, S.E., 1997. Interrenal stress responsiveness of tilapia (Oreochromis mossambicus) is impaired by dietary exposure to PCB 126. Gen. Comp. Endocrinol. 108, 472–82. Radder, R.S., 2007. Maternally derived egg yolk steroid hormones and sex determination: Review of a paradox in reptiles. J. Biosci. 32, 1213–1220. Rahmatullah, M., Boyde, T.R.C., 1980. Improvements in the determination of urea using diacetyl monoxime; methods with and without deproteinisation. Clin. Chim. Acta 107, 3–9. Rajagopalan, K. V, Fridovich, I., Handler, P., 1961. Competitive inhibition of enzyme activity by urea. J. Biol. Chem. 236, 1059–1065. Read, L.J., 1968. Urea and trimethylamine oxide levels in elasmobranch embryos. Biol. Bull. 135, 537–547. Reijnders, P.J.H., 1986. Reproductive failure in common seals feeding on fish from polluted coastal waters. Nature 324, 456–457. Rice, C.D., Schlenk, D., 1995. Immune function and cytochrome P4501A activity after acute exposure to 3, 3’, 4, 4’, 5-pentachlorobiphenyl (PCB 126) in channel catfish. J. Aquat. Anim. Health 7, 195–204. Rivarola, M.A., Forest, M., Migeon, C.J., 1968. Testosterone, androstenedione and dehydroepiandrosterone in plasma during pregnancy and at delivery: concentration and protein binding. J. Clin. Endocrinol. Metab. 28, 34–40. Ross, P.S., Ellisà, G.M., Ikonomou, M.G., Barrett-Lennard, L.G., Addison, R.F., 2000. High PCB concentrations in free-ranging Pacific Killer Whales, Orcinus orca: effects of age, sex and dietary preference. Mar. Pollut. Bull. 40, 504–515. Rossouw, G.J., 1987. Function of the liver and hepatic lipids of the lesser sand shark, Rhinobatos annulatus (Müller & Henle). Comp. Biochem. Physiol. Part B Comp. Biochem. 86, 785– 790. Roy, N.K., Walker, N., Chambers, R.C., Wirgin, I., 2011. Characterization and expression of cytochrome P4501A in Atlantic sturgeon and shortnose sturgeon experimentally exposed to coplanar PCB 126 and TCDD. Aquat. Toxicol. 104, 23–31. Safe, S., Bandiera, S., Sawyer, T., Robertson, L., Safe, L., Parkinson, A., Thomas, P.E., Ryan, D.E., Reik, L.M., Levin, W., Denomme, M.A., Fujita, T. 1985. PCBs: structure–function relationships and mechanism of action. Environ. Health Perspect. 60, 47–56. Safe, S.H., 1994. Polychlorinated biphenyls (PCBs): environmental impact, biochemical and toxic responses, and implications for risk assessment. Crit. Rev. Toxicol. 24, 87–149. Salice, C.J., Rowe, C.L., Eisenreich, K.M., 2014. Integrative demographic modeling reveals population level impacts of PCB toxicity to juvenile snapping turtles. Environ. Pollut. 184,

156

154–160. Sawyna, J.M., Spivia, W.R., Radecki, K., Fraser, D.A., Lowe, C.G., 2017. Association between chronic organochlorine exposure and immunotoxicity in the round stingray (Urobatis halleri). Environ. Pollut. 223, 42–50. Schmidt-Nielsen, B., Truniger, B., Rabinowitz, L., 1972. Sodium-linked urea transport by the renal tubule of the spiny dogfish Squalus acanthias. Comp. Biochem. Physiol. Part A Physiol. 42, 13–25. Schmidt-Nielsen, K., 1997. Animal physiology: adaptation and environment. Cambridge University Press. Schreck, C.B., 2000. Accumulation and long-term effects of stress in fish, in: Moberg, G.P., Mench, J.A. (Eds.), The Biology of Animal Stress. CABI publishing Wallingford, UK, pp. 147–158. Schwabl, H., 1996. Maternal testosterone in the avian egg enhances postnatal growth. Comp. Biochem. Physiol. Part A Physiol. 114, 271–276. Schwacke, L.H., Zolman, E.S., Balmer, B.C., De Guise, S., George, R.C., Hoguet, J., Hohn, A.A., Kucklick, J.R., Lamb, S., Levin, M., Litz, J.A., McFee, W.E., Place, N.J., Townsend, F.I., Wells, R.S., Rowles, T.K., 2011. Anaemia, hypothyroidism and immune suppression associated with polychlorinated biphenyl exposure in bottlenose dolphins (Tursiops truncatus). Proc. R. Soc. London B Biol. Sci. rspb20110665. Serdar, B., LeBlanc, W.G., Norris, J.M., Dickinson, L.M., 2014. Potential effects of polychlorinated biphenyls (PCBs) and selected organochlorine pesticides (OCPs) on immune cells and blood biochemistry measures: a cross-sectional assessment of the NHANES 2003-2004 data. Environ. Heal. 13, 114. Silberhorn, E.M., Glauert, H.P., Robertson, L.W., 1990. Critical Reviews in: Carcinogenicity of Polyhalogenated Biphenyls: PCBs and PBBs. Crit. Rev. Toxicol. 20, 440–496. Singer, T.D., Ballantyne, J.S., 1989. Absence of extrahepatic lipid oxidation in a freshwater elasmobranch, the dwarf stingray Potamotrygon magdalenae: evidence from enzyme activities. J. Exp. Zool. Part A Ecol. Genet. Physiol. 251, 355–360. Skomal, G.B., Mandelman, J.W., 2012. The physiological response to anthropogenic stressors in marine elasmobranch fishes: A review with a focus on the secondary response. Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 162, 146–155. Smith, H.W., 1936. The retention and physiological role of urea in the elasmobranchii. Biol. Rev 11, 49–82. Smith, P.K., Krohn, R. Il, Hermanson, G.T., Mallia, A.K., Gartner, F.H., Provenzano, Md., Fujimoto, E.K., Goeke, N.M., Olson, B.J., Klenk, D.C., 1985. Measurement of protein using bicinchoninic acid. Anal. Biochem. 150, 76–85. Smolders, R., Coen, W. De, Blust, R., 2005. Integrative measures of toxicant exposure in zebra fish (Danio rerio) at different levels of biological organization. Tech. Aquat. Toxicol. 2, 3. Sparling, D.W., Linder, G., Bishop, C.A., Krest, S.K., 2010. Ecotoxicology of amphibians and reptiles, 2nd ed. CRC Press. Spencer, T.E., Sandra, O., Wolf, E., 2008. Genes involved in conceptus--endometrial interactions in ruminants: insights from reductionism and thoughts on holistic approaches. Reproduction 135, 165–179. Spieler, R.E., Fahy, D.P., Sherman, R.L., Sulikowski, J.A., Quinn, T.P., 2013. The Yellow Stingray, Urobatis jamaicensis (Chondrichthyes: Urotrygonidae): a synoptic review 47, 67– 97.

157

Steele, S.L., Yancey, P.H., Wright, P.A., 2005. The Little Skate Raja erinacea exhibits an extrahepatic ornithine urea cycle in the muscle and modulates nitrogen metabolism during low-salinity challenge. Physiol. Biochem. Zool. 78, 216–226. doi:10.1086/427052 Steele, S.L., Yancey, P.H., Wright, P.A., 2004. Dogmas and controversies in the handling of nitrogenous wastes: osmoregulation during early embryonic development in the marine little skate Raja erinacea; response to changes in external salinity. J. Exp. Biol. 207, 2021– 2031. Steinberger, A., Stein, E.D., 2003. Effluent discharges to the Southern California Bight from large municipal wastewater treatment facilities in 2001 and 2002. South. Calif. Coast. Water Res. Proj. Bienn. Rep. 2004, 2–15. Stevens, J.D., Bonfil, R., Dulvy, N.K., Walker, P.A., 2000. The effects of fishing on sharks, rays, and chimaeras (chondrichthyans), and the implications for marine ecosystems. ICES J. Mar. Sci. J. du Cons. 57, 476–494. Stone, B.A., Seamark, R.F., Kelly, R.W., Deam, S., 1986. Production of steroids and release of prostaglandins by spherical pig blastocysts in vitro. Aust. J. Biol. Sci. 39, 283–294. Swedenborg, E., Rüegg, J., Mäkelä, S., Pongratz, I., 2009. Endocrine disruptive chemicals: mechanisms of action and involvement in metabolic disorders. J. Mol. Endocrinol. 43, 1– 10. Tabuchi, M., Veldhoen, N., Dangerfield, N., Jeffries, S., Helbing, C.C., Ross, P.S., 2006. PCB- related alteration of thyroid hormones and thyroid hormone receptor gene expression in free-ranging Harbor Seals (Phoca vitulina). Environ. Health Perspect. 114, 1024–1031. Team, R.D.C., 2011. R: a language and environment for statistical computing. Thomas, P., 1989. Effects of Aroclor 1254 and cadmium on reproductive endocrine function and ovarian growth in atlantic croaker. Mar. Environ. Res. 28, 499–503. Thompson, M.B., Speake, B.K., 2002. Energy and nutrient utilisation by embryonic reptiles. Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 133, 529–538. Thorson, T.B., Gerst, J.W., 1972. Comparison of some parameters of serum and uterine fluid of pregnant, viviparous sharks (Carcharhinus leucas) and serum of their near-term young. Comp. Biochem. Physiol. Part A Physiol. 42, 33–40. Tonack, S., Kind, K., Thompson, J.G., Wobus, A.M., Fischer, B., Santos, A.N., 2007. Dioxin affects glucose transport via the arylhydrocarbon receptor signal cascade in pluripotent embryonic carcinoma cells. Endocrinology 148, 5902–5912. Treberg, J.R., Driedzic, W.R., 2006. Maintenance and accumulation of trimethylamine oxide by winter skate (Leucoraja ocellata): reliance on low whole animal losses rather than synthesis. Am. J. Physiol. Regul. Integr. Comp. Physiol. 291, R1790--R1798. Treberg, J.R., Driedzic, W.R., 2002. Elevated levels of trimethylamine oxide in deep-sea fish: evidence for synthesis and intertissue physiological importance. J. Exp. Zool. 293, 39–45. Treberg, J.R., Speers-Roesch, B., Piermarini, P.M., Ip, Y.K., Ballantyne, J.S., Driedzic, W.R., 2006. The accumulation of methylamine counteracting solutes in elasmobranchs with differing levels of urea: a comparison of marine and freshwater species. J. Exp. Biol. 209, 860–870. Tricas, T.C., Maruska, K.P., Rasmussen, L.E.L., 2000. Annual cycles of steroid hormone production, gonad development, and reproductive behavior in the Atlantic Stingray. Gen. Comp. Endocrinol. 118, 209–225. Trivers, R.L., 1974. Parent-offspring conflict. Am. Zool. 14, 249–264. Van den Berg, M., Birnbaum, L.S., Denison, M., De Vito, M., Farland, W., Feeley, M., Fiedler,

158

H., Hakansson, H., Hanberg, A., Haws, L., Rose, M., Safe, S., Schrenk, D., Tohyama, C., Tritscher, A., Tuomisto, J., Tysklind, M., Walker, N., Peterson, R.E., 2006. The 2005 World Health Organization reevaluation of human and mammalian toxic equivalency factors for dioxins and dioxin-like compounds. Toxicol. Sci. 93, 223–241. Vaudo, J.J., Lowe, C.G., 2006. Movement patterns of the round stingray Urobatis halleri (Cooper) near a thermal outfall. J. Fish Biol. 68, 1756–1766. Vidal-Dorsch, D.E., Bay, S.M., Ribecco, C., Sprague, L.J., Angert, M., Ludka, C., Ricciardelli, E., Carnevali, O., Greenstein, D.J., Schlenk, D., Kelley, K.M., Reyes, J.A., Snyder, S., Vanderford, B., Wiborg, L.C., Petschauer, D., Sasik, R., Baker, M., Hardiman, G., 2013. Genomic and phenotypic response of hornyhead turbot exposed to municipal wastewater effluents. Aquat. Toxicol. 140–141, 174–184. Vijayan, M.M., Aluru, N., Leatherland, J.F. 2010. Stress response and the role of cortisol, in: Fish Diseases and Disorders. CAB International Wallingford, pp. 182–201. Vijayan, M.M., Aluru, N., Maule, A.G., Jørgensen, E.H., 2006. Fasting augments PCB impact on liver metabolism in anadromous Arctic Char. Toxicol. Sci. 91, 431–439. Vijayan, M.M., Pereira, C., Grau, E.G., Iwama, G.K., 1997. Metabolic responses associated with confinement stress in tilapia: the role of cortisol. Comp. Biochem. Physiol. Part C Pharmacol. Toxicol. Endocrinol. 116, 89–95. Walters, D.M., Mills, M.A., Cade, B.S., Burkard, L.P., 2011. Trophic Magnification of PCBs and Its Relationship to the Octanol−Water Partition Coefficient. Environ. Sci. Technol. 45, 3917–3924. doi:10.1021/es103158s Wania, F., Mackay, D., 1996. Tracking the distribution of persistent organic pollutants. Environ. Sci. Technol. 30, 390–396. Ward, P., Myers, R.A., 2005. Shifts in open-ocean fish communities coinciding with the commencement of commerical fishing. Ecology 86, 835–847. Watson, R.R., Dickson, K.A., 2001. Enzyme activities support the use of liver lipid–derived ketone bodies as aerobic fuels in muscle tissues of active sharks. Physiol. Biochem. Zool. 74, 273–282. Wekell, J.C., Barnett, H., 1991. New method for analysis of trimethylamine oxide using ferrous sulfate and EDTA. J. Food Sci. 56, 132–135. Wilson, J.D., Griffin, J.E., George, F.W., 1980. Sexual differentiation: early hormone synthesis and action. Biol. Reprod. 22, 9–17. Wiseman, S., Vijayan, M.M., 2011. Aroclor 1254 disrupts liver glycogen metabolism and enhances acute stressor-mediated glycogenolysis in rainbow trout. Comp. Biochem. Physiol. Part C Toxicol. Pharmacol. 154, 254–260. Wood, S., 2000. Wetlands of the Los Angeles River watershed: profiles and restoration opportunities. California Conservancy. p 142. Wourms, J.P., 1977. Reproduction and development in chondrichthyan fishes. Am. Zool. 17, 379–410. Wourms, J.P., Demski, L.S., 1993. The reproduction and development of sharks, skates, rays and ratfishes: introduction, history, overview, and future prospects, in: The Reproduction and Development of Sharks, Skates, Rays and Ratfishes. Springer, pp. 7–21. Wourms, J.P., Grove, B.D., Lombardi, J., 1988. The maternal-embryonic relationship in viviparous fishes, in: Physiology, W.S.H. and D.J.R.B.T.-F. (Ed.), The Physiology of Developing FishViviparity and Posthatching Juveniles. Academic Press, pp. 1–134. Yancey, P.H., Somero, G.N., 1980. Methylamine osmoregulatory solutes of elasmobranch fishes

159

counteract urea inhibition of enzymes. J. Exp. Zool. 212, 205–213. Yancey, P.H., Somero, G.N., 1979. Counteraction of urea destabilization of protein structure by methylamine osmoregulatory compounds of elasmobranch fishes. Biochem. J. 183, 317– 323. Ylitalo, G.M., Stein, J.E., Hom, T., Johnson, L.L., Tilbury, K.L., Hall, A.J., Rowles, T., Greig, D., Lowenstine, L.J., Gulland, F.M.D., 2005. The role of organochlorines in cancer- associated mortality in California sea lions (Zalophus californianus). Mar. Pollut. Bull. 50, 30–39. Zambrano, E., Guzmán, C., Rodríguez-González, G.L., Durand-Carbajal, M., Nathanielsz, P.W., 2014. Fetal programming of sexual development and reproductive function. Mol. Cell. Endocrinol. 382, 538–549. Zhou, T., Scali, R., Weis, J.S., 2001. Effects of methylmercury on ontogeny of prey capture ability and growth in three populations of larval Fundulus heteroclitus. Arch. Environ. Contam. Toxicol. 41, 47–54.