<%BANNER%>

Reproductive and Stress Analyses in West Indian Manatees (Trichechus manatus)

Permanent Link: http://ufdc.ufl.edu/UFE0021927/00001

Material Information

Title: Reproductive and Stress Analyses in West Indian Manatees (Trichechus manatus)
Physical Description: 1 online resource (135 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: acth, corpus, cortisol, manatee, ovary, pregnancy, progesterone, reproduction, star, tem
Veterinary Medicine -- Dissertations, Academic -- UF
Genre: Veterinary Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: A high sensitivity chemiluminescent serum progesterone assay was validated for pregnancy diagnosis in manatees, because an endocrine-based diagnostic was not previously available. A diagnostic pregnancy threshold of 0.4 ng/ml was determined and repeat sampling (?2x) is recommended to confirm pregnancy. A value of 0.1 ng/ml was indicative of impending abortion in two late pregnancy females. Gas chromatography-mass spectrometry was used to test for progesterone metabolites in manatee plasma, including the 5?-reduced progestins (predominant during elephant gestation), 17?-hydroxyprogesterone (common in Asian elephants), and 20?-OHP (an inactive metabolite). It was found that progesterone (pregn-4-ene-3,20-dione) itself predominates in manatee plasma. Manatee corpora lutea were morphologically characterized and found to be significantly larger in early diestrus (n=22) and mid pregnancy (n=8) than late pregnancy (n=66) (P < 0.05). Granulosa lutein cells (GLCs) predominated, were significantly larger in early diestrus (max=60.8 ?m; n=234) than late pregnancy (max=37.5 ?m; n=576), and were significantly larger than theca lutein cells (n=1,015). Ultrastructurally, degenerating corpora lutea were characterized by a lack of organization, few identifiable organelles, collagen infiltration, and occlusion of capillaries by distended endothelial cells. The steroidogenic potential of manatee follicles and corpora lutea was determined using immunohistochemical staining with steroidogenic acute regulatory protein (StAR). Expression of StAR in manatee follicles (n=131) was significantly greater during early diestrus (n=23) than anestrus (n=24) or pregnancy (n=84), while expression in corpora lutea was significantly greater in early diestrus (n=19) and late pregnancy (n=15) than abortion (n=4). Methods were developed to measure manatee stress, particularly because stress may affect reproduction. Hydrocortisone (cortisol) was the predominant manatee glucocorticoid, and chemiluminescent assays for cortisol and ACTH (adrenocorticotropic hormone) were validated. Cortisol concentrations ?1.0 ?g/dl were diagnostic of chronic stress while ACTH concentrations ?87.5 pg/ml were diagnostic of peracute stress. Concentrations of ACTH were positively correlated with capture time (P=0.00686) and lactate concentrations (P=0.0000343). Our study resulted in a pregnancy diagnostic, identification of circulating progestins, description of corpus luteum morphology, determination of the steroidogenic potential of follicles and corpora lutea, and identification of important stress indicators. This information will benefit manatee biologists, veterinarians, and managers, and help optimize management of this endangered marine mammal.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Samuelson, Don A.
Local: Co-adviser: Verstegen, John P. L.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021927:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021927/00001

Material Information

Title: Reproductive and Stress Analyses in West Indian Manatees (Trichechus manatus)
Physical Description: 1 online resource (135 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: acth, corpus, cortisol, manatee, ovary, pregnancy, progesterone, reproduction, star, tem
Veterinary Medicine -- Dissertations, Academic -- UF
Genre: Veterinary Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: A high sensitivity chemiluminescent serum progesterone assay was validated for pregnancy diagnosis in manatees, because an endocrine-based diagnostic was not previously available. A diagnostic pregnancy threshold of 0.4 ng/ml was determined and repeat sampling (?2x) is recommended to confirm pregnancy. A value of 0.1 ng/ml was indicative of impending abortion in two late pregnancy females. Gas chromatography-mass spectrometry was used to test for progesterone metabolites in manatee plasma, including the 5?-reduced progestins (predominant during elephant gestation), 17?-hydroxyprogesterone (common in Asian elephants), and 20?-OHP (an inactive metabolite). It was found that progesterone (pregn-4-ene-3,20-dione) itself predominates in manatee plasma. Manatee corpora lutea were morphologically characterized and found to be significantly larger in early diestrus (n=22) and mid pregnancy (n=8) than late pregnancy (n=66) (P < 0.05). Granulosa lutein cells (GLCs) predominated, were significantly larger in early diestrus (max=60.8 ?m; n=234) than late pregnancy (max=37.5 ?m; n=576), and were significantly larger than theca lutein cells (n=1,015). Ultrastructurally, degenerating corpora lutea were characterized by a lack of organization, few identifiable organelles, collagen infiltration, and occlusion of capillaries by distended endothelial cells. The steroidogenic potential of manatee follicles and corpora lutea was determined using immunohistochemical staining with steroidogenic acute regulatory protein (StAR). Expression of StAR in manatee follicles (n=131) was significantly greater during early diestrus (n=23) than anestrus (n=24) or pregnancy (n=84), while expression in corpora lutea was significantly greater in early diestrus (n=19) and late pregnancy (n=15) than abortion (n=4). Methods were developed to measure manatee stress, particularly because stress may affect reproduction. Hydrocortisone (cortisol) was the predominant manatee glucocorticoid, and chemiluminescent assays for cortisol and ACTH (adrenocorticotropic hormone) were validated. Cortisol concentrations ?1.0 ?g/dl were diagnostic of chronic stress while ACTH concentrations ?87.5 pg/ml were diagnostic of peracute stress. Concentrations of ACTH were positively correlated with capture time (P=0.00686) and lactate concentrations (P=0.0000343). Our study resulted in a pregnancy diagnostic, identification of circulating progestins, description of corpus luteum morphology, determination of the steroidogenic potential of follicles and corpora lutea, and identification of important stress indicators. This information will benefit manatee biologists, veterinarians, and managers, and help optimize management of this endangered marine mammal.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Samuelson, Don A.
Local: Co-adviser: Verstegen, John P. L.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021927:00001


This item has the following downloads:


Full Text

PAGE 1

REPRODUCTIVE AND STRESS ANALYSES IN WEST INDIAN MANATEES (Trichechus manatus) By KATHLEEN M. TRIPP A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Kathleen M. Tripp 2

PAGE 3

To mom, Rich, Hannah, and Monte. 3

PAGE 4

ACKNOWLEDGMENTS Samples obtained in the United States were collected under the following permit numbers from the U.S. Fish and Wildlife Service: MA773494-8 (Florida Fish and Wildlife Conservation Commission Fish and Wildlife Research Institute) and MA791721 (U.S. Geological Survey Sirenia Project). Mexico samples were collected under permits NUM/SGPA/DGVS/03144 and 04513-2005 from the Direcion General del Vida Silvestre and transported to the United States under CITES export permit MX24463 (2005) and import permit US808447/9. Manatee plasma samples were shipped to Belgium for GC/MS/MS analysis using Belgium CITES import Number 705/2007. The Mote Marine Lab study was completed under Institutional Animal Care and Use Committee approval from that institution (registration number: 06-10-KT1). Samples were examined at the University of Florida under U.S. Fish and Wildlife Service permit MA067116-0 (Aquatic Animal Health Program). Research protocols were approved by the University of Florida Institutional Animal Care and Use Committee (D956). I gratefully acknowledge the University of Floridas Aquatic Animal Health Program for funding this research. I must also thank my committee members for their mentorship Drs. Don Samuelson, John Verstegen, Kendal Harr, Michael Fields, and Barbara Sheppard. Thanks also go to Pat Lewis, Jim Burrows, and Melanie Pate, for their technical expertise in the laboratory, and to Lourdes Gonzalez for her assistance with the morphometry study. This project would not have been possible without the cooperation of numerous agencies and aquaria, including Bob Bonde and Susan Butler from the U.S. Geological Survey Sirenia Project; Martine DeWit, Martha Keller, Ken Arrison, Chip Deutsch, Margie Barlas, and Kari Higgs from the Fish and Wildlife Research Institute; Benjamin Morales from Ecosur Chetumal; and the capture teams of these institutions. I thank the veterinarians and staff at the Cincinnati, Columbus, and Lowry Park Zoos and to those at the Homosassa Springs State Wildlife Park, 4

PAGE 5

Miami Seaquarium, Mote Marine Laboratory, Sea World of Florida, and The Seas, with special thanks to Dr. Maya Rodriguez, Dr. Charles Manire, Joe Gaspard, and Dr. Elizabeth Nolan. I thank Dr. Dennis Schmitt of Missouri State University for collaboration on methods development, Steve Lamb of Cornell Universitys Animal Health Diagnostic Lab for methods comparison and corticosterone testing, the CER Endocrinology Laboratory in Belgium for GG/MS/MS analysis, and Dr. John Harvey of the University of Florida for clinical data. Finally, thanks go to Nicole Adimey of the U.S. Fish and Wildlife Service for facilitating sample collection. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES .........................................................................................................................10 LIST OF FIGURES .......................................................................................................................12 LIST OF ABBREVIATIONS ........................................................................................................13 ABSTRACT ...................................................................................................................................16 CHAPTER 1 INTRODUCTION ..................................................................................................................18 CHAPTER 2 VALIDATION OF A SERUM IMMUNOASSAY TO MEASURE PROGESTERONE AND DIAGNOSE PREGNANCY IN THE WEST INDIAN MANATEE ...........................22 Introduction .............................................................................................................................22 Materials and Methods ...........................................................................................................24 Blood Sampling ...............................................................................................................24 Sample Population ...........................................................................................................24 Laboratory Analysis ........................................................................................................27 Statistical Analysis ..........................................................................................................28 Results .....................................................................................................................................29 Assay Analytical and Diagnostic Validation ...................................................................29 Effects of Freeze-Thaw ...................................................................................................30 Normal Gestational Changes ...........................................................................................31 Abortion Detection ..........................................................................................................31 Seasonal Variation ...........................................................................................................31 Progesterone Concentrations by Gender and Age Class .................................................32 Discussion ...............................................................................................................................32 Assay Analytical and Diagnostic Validation ...................................................................32 Effects of Freeze-Thaw ...................................................................................................33 Normal Gestational Changes ...........................................................................................33 Abortion Detection ..........................................................................................................34 Seasonal Variation ...........................................................................................................35 Progesterone Concentrations by Gender and Age Class .................................................36 Future Studies ..................................................................................................................36 Conclusion .......................................................................................................................37 6

PAGE 7

CHAPTER 3 DETECTION AND IDENTIFICATION OF PROGESTERONE METABOLITES IN THE FLORIDA MANATEE USING GC/MS/MS ................................................................41 Introduction .............................................................................................................................41 Materials and Methods ...........................................................................................................42 Blood Sampling ...............................................................................................................42 Sample Population ...........................................................................................................43 Laboratory Analysis ........................................................................................................43 Data Analysis ...................................................................................................................45 Results .....................................................................................................................................45 Identification of the Predominant Progestin ....................................................................45 Gender and Reproductive Differences in Progestins .......................................................46 Discussion ...............................................................................................................................47 Identification of the Predominant Progestin ....................................................................47 Gender and Reproductive Differences in Progestins .......................................................50 Future Studies ..................................................................................................................51 Conclusion .......................................................................................................................52 CHAPTER 4 MORPHOLOGICAL CHARACTERIZATION OF CORPORA LUTEA AND LUTEAL CELLS IN THE FLORIDA MANATEE ...............................................................55 Introduction .............................................................................................................................55 Materials and Methods ...........................................................................................................57 Sample Collection ...........................................................................................................57 Sample Population ...........................................................................................................57 Sample Processing ...........................................................................................................58 Morphometric and TEM Analysis ...................................................................................59 Statistical Analysis .........................................................................................................60 Results .....................................................................................................................................61 The Manatee Corpus Luteum ..........................................................................................61 Luteal Cells ......................................................................................................................62 Endothelial Cells and Fibroblasts ....................................................................................64 Degenerative Changes Observed with TEM ...................................................................66 Discussion ...............................................................................................................................67 The Manatee Corpus Luteum ..........................................................................................67 Luteal Cells ......................................................................................................................69 Endothelial Cells and Fibroblasts ....................................................................................73 Degenerative Changes Observed with TEM ...................................................................74 Conclusion .......................................................................................................................75 7

PAGE 8

CHAPTER 5 ASSESSMENT OF THE STEROIDOGENIC POTENTIAL OF FLORIDA MANATEE FOLLICLES AND CORPORA LUTEA USING STEROIDOGENIC ACUTE REGULATORY PROTEIN ...................................................................................................85 Introduction .............................................................................................................................85 Materials and Methods ...........................................................................................................86 Sample Collection ...........................................................................................................86 Sample Population ...........................................................................................................86 Immunohistochemistry ....................................................................................................87 Staining Analysis .............................................................................................................88 Statistical Analysis ..........................................................................................................88 Results .....................................................................................................................................89 Localization of StAR .......................................................................................................89 Intra-Animal Differences .................................................................................................89 Inter-Animal Differences .................................................................................................90 Discussion ...............................................................................................................................91 Localization of StAR .......................................................................................................91 Intra-Animal Differences .................................................................................................92 Inter-Animal Differences .................................................................................................94 Future Studies ..................................................................................................................95 Conclusion .......................................................................................................................95 CHAPTER 6 EVALUATION OF PERACUTE AND CHRONIC STRESS IN THE FLORIDA MANATEE .............................................................................................................................99 Introduction .............................................................................................................................99 Materials and Methods .........................................................................................................101 Blood Sampling .............................................................................................................101 Sample Population .........................................................................................................101 Laboratory Analysis ......................................................................................................103 Statistical Analysis ........................................................................................................105 Results ...................................................................................................................................107 Glucocorticoid Determination .......................................................................................107 Assay Analytical and Diagnostic Validation .................................................................107 Population Baseline and Seasonal Variation .................................................................109 Correlation of ACTH with Capture Parameters ............................................................109 Discussion .............................................................................................................................109 Glucocorticoid Determination .......................................................................................109 Assay Analytical and Diagnostic Validation .................................................................110 Population Baseline and Seasonal Variation .................................................................112 Correlation of ACTH with Capture Parameters ............................................................113 Future Studies ................................................................................................................114 Conclusion .....................................................................................................................114 8

PAGE 9

CHAPTER 7 CONCLUSIONS ..................................................................................................................118 LIST OF REFERENCES .............................................................................................................121 BIOGRAPHICAL SKETCH .......................................................................................................135 9

PAGE 10

LIST OF TABLES Table page 2-1. Manatee Serum Progesterone Analytical Validation Results ................................................37 2-2. Progesterone (P4) Concentrations for Pregnancy Diagnosis throughout Gestation ..............38 2-3. Gestational Progesterone Concentrations Determined from Study Females .........................38 2-4. Abortion-Associated Serum Progesterone (P4) (ng/ml) in Late Pregnant Females ..............38 2-5. Seasonal Progesterone Values (mean P4 SD) for Non-Pregnant and Pregnant Florida Manatees ............................................................................................................................38 2-6. Progesterone Concentrations by Gender and Age Class .......................................................39 3-1. Manatee GC/MS/MS Results for Progesterone and Its Metabolites .....................................53 4-1. Corpora Lutea (CL) Morphometrics for Study Specimens ...................................................76 4-2. Determination of Significant Differences in Mean GLC and TLC Lengths .........................76 4-3. Mean Measurements and Ranges of GLCs ...........................................................................77 4-4. Mean TLC Measurements .....................................................................................................77 4-5. Mean Endothelial Cell Measurements ...................................................................................78 4-6. Endothelial Cell Nuclei Summary: Longitudinal (LS) and Cross Sections (XS) ..................78 4-7. Mean Fibroblast Nuclei Longitudinal Sections .....................................................................79 4-8. Fibroblast Nucleus (Longitudinal Sections) Summary Data .................................................79 5-1. Follicle Summary Data ..........................................................................................................96 5-2. Corpora Lutea (CLs) Summary Data .....................................................................................96 6-1. Cortisol and ACTH Analytical Validation Results .............................................................115 6-2. Cortisol and ACTH Results Associated with Different Age Classes ..................................115 6-3. Serum Cortisol Pairwise Stress Category ComparisonsThresholds Where All ROC Parameters Optimized ......................................................................................................116 6-4. Cortisol and ACTH Descriptive Statistics Among Three Stress Categories .......................116 10

PAGE 11

6-5. Pairwise Stress Category Comparisons for ACTHThresholds Where All ROC Parameters Optimized ......................................................................................................116 6-6. Seasonal Analyte Concentrations in Unstressed Manatees .................................................116 6-7. Parameters Associated with Increasing Follow Time in Three Manatees ...........................117 11

PAGE 12

LIST OF FIGURES Figure page 2-1. Scatter plot for methods comparison. ....................................................................................39 2-2. Progesterone profile generated from a single female of known gestation ............................40 3-1. The molecular structures of progestins investigated in manatee plasma. ..............................54 4-1. Gross sections of manatee corpora lutea ...............................................................................80 4-2. Appearance of GLCs versus a TLC .......................................................................................81 4-3. Micrographs A-F illustrate the microscopic appearance of granulosa lutein and support cells of corpora lutea from each of the studys six females. ..............................................82 4-4. A binucleated granulosa lutein cell from female #1, designated with a star, is surrounded by uninucleated lutein cells. (400x) ................................................................83 4-5. Endothelial cells and fibroblasts are visible in a section of a corpus luteum from female #2 (late diestrus) .................................................................................................................83 4-6. Cross section of a degenerating granulosa lutein cell from female #4 observed with TEM ...................................................................................................................................84 4-7. Higher magnification view of organelles from Figure 4-6. ...................................................84 5-1. A primary follicle from a female in early diestrus stained with StAR (A) and H&E (B). ....96 5-2. Differential StAR staining in the corpora lutea of female #2. ...............................................97 5-3. StAR and corresponding H&E images for 2 pregnant female manatees ..............................98 12

PAGE 13

LIST OF ABBREVIATIONS 5 -DHP 5-pregnane-3,20-dione 5-P3-OH 3-hydroxy-5-pregnan-20-one 17-OHP 17-hydroxyprogesterone 20-OHP 20-hydroxy-4-pregnen-3-one; 20-hydroxyprogesterone ACTH adrenocorticotropic hormone AEC 3-amino, 9 ethyl carbazole chromagen ANOVA Analysis of Variance C degrees Celsius cc cubic centimeter; 1cc=1ml CI confidence interval CL corpus luteum cm centimeter COR cortisol (hydrocortisone) CV coefficient of variation EDTA ethylene diamine tetraacetic acid EtOH ethyl alcohol GC/MS gas chromatography-mass spectrometry GC/MS/MS gas chromatography with tandem quadrupole detectors GLC granulosa lutein cell g gram H&E hematoxylin and eosin H 2 O 2 hydrogen peroxide 13

PAGE 14

HRP horse radish peroxidase i.d. interior diameter kPa kilo Pascal LS longitudinal section LiHep lithium heparin plasma LLOD lower limit of detection m meter M molar min minimum max maximum MeOH methanol mg milligram ml milliliter mmol/L millimoles per liter MRM multiple reaction monitoring MSR molecular stress response MSTFA N-Methyl-N-trifluoroacetamide n sample size NBF neutral buffered formalin ng/ml nanograms per milliliter NPV negative predictive value P P-value; probability of incorrectly rejecting null hypothesis Pa Pascal 14

PAGE 15

P4 progesterone PBS phospho-buffered saline PPV positive predictive value pregn-4-ene-3,20-dione progesterone r correlation coefficient RER rough endoplasmic reticulum RIA radioimmunoassay ROC receiver operating characteristic SAA serum amyloid A SD standard deviation StAR Steroidogenic Acute Regulatory Protein T. Trichechus Td3 16, 16, 17-d 3 -testosterone TEa total allowable error TEM transmission electron microscopy TLC theca lutein cell x g times gravity XS cross section Greek letter alpha; 1-= statistical power for power calculations g/dl micrograms per deciliter g/ml micrograms per milliliter l microliter m micrometer 15

PAGE 16

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy REPRODUCTIVE AND STRESS ANALYSES IN WEST INDIAN MANATEES (Trichechus manatus) By Kathleen M. Tripp May 2008 Chair: Don A. Samuelson Cochair: John P. Verstegen Major: Veterinary Medical Sciences A high sensitivity chemiluminescent serum progesterone assay was validated for pregnancy diagnosis in manatees, because an endocrine-based diagnostic was not previously available. A diagnostic pregnancy threshold of 0.4 ng/ml was determined and repeat sampling (2x) is recommended to confirm pregnancy. A value of 0.1 ng/ml was indicative of impending abortion in two late pregnancy females. Gas chromatography-mass spectrometry was used to test for progesterone metabolites in manatee plasma, including the 5-reduced progestins (predominant during elephant gestation), 17-hydroxyprogesterone (common in Asian elephants), and 20-OHP (an inactive metabolite). It was found that progesterone (pregn-4-ene-3,20-dione) itself predominates in manatee plasma. Manatee corpora lutea were morphologically characterized and found to be significantly larger in early diestrus (n=22) and mid pregnancy (n=8) than late pregnancy (n=66) (P<0.05). Granulosa lutein cells (GLCs) predominated, were significantly larger in early diestrus (max=60.8 m; n=234) than late pregnancy (max=37.5 m; n=576), and were significantly larger than theca lutein cells (n=1,015). Ultrastructurally, degenerating corpora lutea were 16

PAGE 17

characterized by a lack of organization, few identifiable organelles, collagen infiltration, and occlusion of capillaries by distended endothelial cells. The steroidogenic potential of manatee follicles and corpora lutea was determined using immunohistochemical staining with steroidogenic acute regulatory protein (StAR). Expression of StAR in manatee follicles (n=131) was significantly greater during early diestrus (n=23) than anestrus (n=24) or pregnancy (n=84), while expression in corpora lutea was significantly greater in early diestrus (n=19) and late pregnancy (n=15) than abortion (n=4). Methods were developed to measure manatee stress, particularly because stress may affect reproduction. Hydrocortisone (cortisol) was the predominant manatee glucocorticoid, and chemiluminescent assays for cortisol and ACTH (adrenocorticotropic hormone) were validated. Cortisol concentrations 1.0 g/dl were diagnostic of chronic stress while ACTH concentrations 87.5 pg/ml were diagnostic of peracute stress. Concentrations of ACTH were positively correlated with capture time (P=0.00686) and lactate concentrations (P=0.0000343). Our study resulted in a pregnancy diagnostic, identification of circulating progestins, description of corpus luteum morphology, determination of the steroidogenic potential of follicles and corpora lutea, and identification of important stress indicators. This information will benefit manatee biologists, veterinarians, and managers, and help optimize management of this endangered marine mammal. 17

PAGE 18

CHAPTER 1 INTRODUCTION There has been a significant research effort to promote manatee conservation, which includes brief capture of free-ranging manatees for health assessments, rescue of injured or sick individuals, and necropsy of deceased manatees to determine cause of death. Despite this interest in research and conservation, biologists, clinicians, and managers still lack key information regarding reproduction and stress in manatees. Historically, the absence of an early pregnancy diagnostic has caused biologists and clinicians to rely upon visual indications of pregnancy that appear late in the 12 month gestation. These include highly variable and subjective changes such as distension of the abdomen and vulva and have left biologists and clinicians without methods to diagnose or monitor early pregnancy [1]. The absence of an endocrine-based pregnancy diagnostic is due in part to the absence of captive breeding in the United States and concerns related to the capture of pregnant females. Isolation of a specific chorionic gonadotropin, used to diagnose pregnancy in species such as the horse [2,3] has not been attempted. Additionally, while pregnancy-specific hormones including prolactin and relaxin have been identified in species such as African (Loxodonta africana) and Asian elephants (Elephas maximus) [4-6], which are close terrestrial relatives of manatees [7,8], similar diagnostic hormones have not been identified in manatees. Furthermore, abdominal ultrasound, which can be used to diagnose pregnancy in many species [9,10], can only be utilized in manatees after the first 10 weeks of gestation, and its sensitivity is lowered by the manatees expansive gas-filled gastrointestinal tract (M. Rodriguez pers. comm.). In some domestic mammals, pregnancy is associated with progesterone concentrations greater than those observed during non-pregnant diestrus. In the ewe, for example, progesterone concentrations average 0.4 0.4 ng/ml during diestrus and increase to 3.3 0.9 ng/ml during 18

PAGE 19

pregnancy [11]. Low circulating progesterone concentrations have been hypothesized in manatees [12], which would further complicate pregnancy diagnosis. In African and Asian elephants, progesterone (pregn-4-ene-3,20-dione) concentrations during pregnancy are also low compared to many species [10,13-15], and peak below 2.0 ng/ml [16]. These findings led investigators to examine the presence of progesterone metabolites that could be used to maintain pregnancy in elephants. It was found that elephants do in fact possess metabolites, including the 5-reduced progestins in the African elephant [17] and 17-hydroxyprogesterone in the Asian elephant [18]. The 5-reduced metabolites are also predominant during the second half of gestation in the mare [19]. The presence of such metabolites in the peripheral circulation of the manatee has not previously been investigated. In most species, the corpus luteum, a temporary progesterone-producing endocrine organ, develops from an ovulated follicle [20]. Corpora lutea contain at least two types of steroidogenic cells [21], and it is the granulosa lutein cells, characterized by large amounts of smooth endoplasmic reticulum [22] and clusters of mitochondria, which synthesize progesterone [23]. Typically, species that are monotocous (produce one offspring per pregnancy), such as the cow, produces a single corpus luteum [23]. Although manatees are also monotocous (produce one offspring per pregnancy), with only rare occurrences of twinning [24], a large number of follicles and corpora lutea develop within the ovaries. Another monotocous species, the horse, exhibits multiple accessory corpora lutea [25] formed by luteinization of unovulated follicles. The benefit to the manatee of producing a high number of luteal structures is unknown. Also unknown is the steroidogenic contribution of individual corpora lutea and any relationship between this steroidogenic potential and the observed number of corpora lutea. Knowledge of 19

PAGE 20

the structure and function of manatee corpora lutea is required to understand the steroidogenic contribution of these structures during the estrous cycle and pregnancy. The Florida manatee lives in a challenging environment with threats that include watercraft collision, line entanglement, red tide, and cold stress. However, investigations of manatee stress (both peracute and chronic) that may be associated with capture, injury, or disease have been limited [26]. Adrenocorticotropic hormone or ACTH is secreted from the anterior pituitary within minutes of peracute stress exposure and triggers glucocorticoid secretion from the adrenal cortex [27]. The predominant and functional glucocorticoid found in birds and small mammals is corticosterone, while cortisol (hydrocortisone) predominates in medium to large mammals, as well as fish and humans [28]. Circulating ACTH peaks and subsides quickly, while a glucocorticoid peak with a 30-60 minute duration [27] occurs 1 to 4 hours after encountering the peracute stressor [29-34]. The intensity of the stressor determines the degree and duration of the endocrine stress response [27,30]. During the stress response, glucocorticoids help provide energy by upregulating lipolysis and glucose formation, altering metabolic pathways and vasculature via catecholamines, and generally decreasing immune response in order to minimize cell and tissue damage [35,36]. However, long periods of heightened glucocorticoid concentrations may be detrimental as they shift energy away from other important biological processes. Problems associated with prolonged glucocorticoid increase include inhibition of reproductive cycles, abortion, and death [27,29,37]. There is disagreement over the level and duration of cortisol increase that can be harmful, but most studies agree that more than several days of stress can begin to cause physiological effects [28,30,38,39]. The absence of important information related to manatee reproduction and stress led to the current study, with the following objectives: 20

PAGE 21

Objective 1: To develop a progesterone-based pregnancy diagnostic. Objective 2: To determine the presence of progesterone metabolites in plasma. Objective 3: To morphologically describe corpora lutea using light microscopy and transmission electron microscopy. Objective 4: To determine the steroidogenic potential of corpora lutea. Objective 5: To establish methods for evaluating peracute and chronic stress. Our study provided an endocrine-based pregnancy diagnostic for the manatee and identified progesterone metabolites present during diestrus and pregnancy. Additionally, changes to manatee corpus luteum morphology during diestrus and pregnancy were described and the steroidogenic potential of follicles and corpora lutea was determined. Finally, important stress thresholds were identified, which can be used to guide management of this endangered species. Each of the objectives stated above corresponds to one chapter of the dissertation. These chapters are written as stand-alone manuscripts for publication. 21

PAGE 22

CHAPTER 2 VALIDATION OF A SERUM IMMUNOASSAY TO MEASURE PROGESTERONE AND DIAGNOSE PREGNANCY IN THE WEST INDIAN MANATEE Introduction There is minimal peer-reviewed literature regarding serum progesterone concentrations in manatees and endocrine-based pregnancy diagnosis has not previously been possible. This is due in part to the absence of captive breeding in the United States and concerns related to the capture of pregnant females. Isolation of a specific chorionic gonadotropin as occurs in the horse [2,3] has not been attempted. Additionally, while pregnancy-specific hormones including prolactin and relaxin have been identified in species such as African and Asian elephants [4-6], which are close terrestrial relatives of manatees [7,8], similar diagnostic hormones have not been identified in manatees. Furthermore, abdominal ultrasound, which can be used to diagnose pregnancy in many species [9,10], can only be utilized in manatees after the first 10 weeks of gestation, and its sensitivity is lowered by the manatees expansive, gas-filled gastrointestinal tract (M. Rodriguez pers. comm.). Historically, the absence of an early pregnancy diagnostic has caused biologists and clinicians to rely upon visual indications of pregnancy that appear late in the 12 month gestation [1]. These include highly variable and subjective parameters such as distension of the abdomen and vulva and have left biologists and clinicians without methods to diagnose or monitor early pregnancy [1]. However, the ability to diagnose pregnancy is valuable, particularly for rehabilitating manatees undergoing active medical treatment for such traumatic injuries as watercraft strike and entanglement, because treatment may be altered for pregnant females. Earlier studies measured serum progesterone in captive manatees [40] and utilized fecal samples to estimate the duration of the manatee estrous cycle as 28 to 42 days [12]. In the first study, serum progesterone in a captive manatee was >3.0 ng/ml [40], while the second study 22

PAGE 23

found that progesterone concentrations in pregnant females exhibited a wide range, wherein lower values could overlap with those of non-pregnant females and males [12]. While fecal hormone data provide estimates of estrous cycle length and progesterone concentrations, the use of fecal steroid assays is complicated by factors related to sample collection and storage and steroid extraction [41]. The manatees long gut transit time (approximately 7 days) [42] further contributes to the imprecision of manatee fecal steroid hormone data, making it less suitable for clinical pregnancy diagnosis. In many domestic mammals, pregnancy is associated with progesterone concentrations greater than those observed during non-pregnant diestrus. In the ewe, for example, progesterone concentrations in non-pregnant females average 0.4 0.4 ng/ml and increase to 3.3 0.9 ng/ml during pregnancy [11]. Progesterone concentrations are also higher during pregnancy for goats [10], cows [13], alpacas [14], and sows [15]. In Asian and African elephants, progesterone concentrations during pregnancy are low in comparison with those of many other species. Nevertheless, most elephant studies have found that concentrations are still greater during pregnancy than diestrus [6,43], although serial sampling is recommended to confirm pregnancy [44]. This observation of low but diagnostic serum progesterone concentrations during elephant pregnancy [6,43] suggested that a high sensitivity serum progesterone assay may be useful for diagnosing manatee pregnancy. The objectives of this study were to: (1) analytically and diagnostically validate a chemiluminescent serum progesterone assay for pregnancy diagnosis in healthy manatees; (2) determine the effect of repeated freeze-thaw cycles on progesterone results; (3) document normal changes to progesterone concentration through gestation; (4) identify the progesterone concentration indicative of abortion; (5) identify any seasonal progesterone variation in 23

PAGE 24

non-pregnant and pregnant females; and (6) examine differences in progesterone concentration throughout the manatee population, considering gender, age class, and reproductive status. This information will allow biologists, clinicians, and managers to better assess the reproductive status of free-ranging and captive manatees. Materials and Methods Blood Sampling Free-ranging manatees were captured using nets deployed from boats [45]. Blood was collected from the brachial vascular bundle during all seasons of the year as part of ongoing, government-led ecological investigations [46-49]. Such health assessments were conducted for free-ranging and captive Florida manatees (T. manatus latirostris), as well as free-ranging Antillean manatees (T. manatus manatus) in Mexico and Puerto Rico. Serum was separated within 1 hour of collection by onsite centrifugation (3,000 x g for 10 min.). Samples were then transported on ice and either assayed immediately upon arrival at the University of Florida or stored at -80C for later analysis. Sample Population Serum samples from 114 manatees were analyzed, including females (n=69), males (45), calves (n=9), subadults (n=33), and adults (n=72). Thirteen replicate samples for each of three females were used to test intra-assay precision (repeated specimen testing on the same day), and three replicate samples from six manatees were used to assess inter-assay precision (repeated specimen testing in different months). For accuracy, dilution experiments were used to test parallelism and recoverability of progesterone using samples from three manatees. A methods comparison was completed using paired samples from 40 manatees. Only adult females, defined by a total straight length >265 cm [50], were included in the sample population for diagnostic validation (determination of the progesterone concentration that 24

PAGE 25

distinguishes non-pregnancy and pregnancy) using receiver operating characteristic (ROC) analysis [51]. Additionally, healthy females were used to establish the reference interval for the non-pregnant population because inflammation associated with disease may affect progesterone concentrations [52]. Health was assessed by field observations (including body condition, heart rate, and respiratory rate) and standard biochemical analysis of serum amyloid A (SAA) [53]. In the pregnant manatee group, three females with SAA >70 g/ml (and 130 g/ml) were included as their inflammatory processes were not severe enough to interfere with delivery of healthy calves [54]. No long-term captive females were utilized for diagnostic validation, to minimize confounding factors associated with numerous infertile estrous cycles, as has been documented in elephants [55]. Healthy females who were lactating and observed with calves at the time of their capture were included in the non-pregnant population. Pregnancy was confirmed using physical signs or presence of a newborn calf that indicated an individual had been pregnant on the date of her previous blood sampling. Healthy manatees that had been in a rehabilitation or temporary captive setting were assessed as non-pregnant once they had been in captivity for >7 months without showing signs of pregnancy or abortion. Additionally, subsequent field observations of previously captured, free-ranging females, showing no signs of pregnancy or a dependent calf at approximately 1 year post-capture, were used to assess females as non-pregnant. This method of pregnancy diagnosis cannot account for false negative assessment due to abortion or perinatal mortality within the free-ranging population. The first ROC analysis compared progesterone concentrations in non-pregnant females (n=30) and pregnant females throughout gestation (n=15). Two additional ROC analyses compared progesterone values in non-pregnant females (n=30) with values from females in the 25

PAGE 26

first (months 1 to 6; n=5) and second halves of gestation (months 7 to 12; n=10). Serum samples from nine of these female manatees were used to test the effect of freeze-thaw on manatee serum progesterone results. A single early gestation rehabilitating manatee was sampled for blood 13 times over a 6 month time period until pregnancy was diagnosed. This female was not known to be pregnant at the time of her capture, and once physical signs of pregnancy appeared, sampling was stopped to minimize danger to the fetus. Additionally, single samples from other females with known gestation dates were used to examine trends in progesterone concentration during early (months 1 to 4; n=9), mid (months 5 to 8; n=8), and late (months 9 to 12; n=5) pregnancy. Progesterone concentrations associated with abortion were defined using samples available from three late gestation manatees known to have aborted a fetus during their treatment for injuries at a rehabilitation facility. Manatee calves may be born ranging in length from 82 to 160 cm or 95 to 155 cm [1,24], indicating that these fetuses, which were >81 cm in total length, were in the final months of gestation. Progesterone concentrations for female Florida manatees were evaluated for seasonal variation, since the Florida manatee has been identified as a diffusely seasonal breeder [1]. Seasonal variation in progesterone concentrations was examined for Florida non-pregnant (n=21) and pregnant females (n=14). For non-pregnant females, samples were available from winter (n=3), spring (n=1), summer (n=5), and autumn (n=12). Samples from pregnant females also included winter (n=4), spring (n=2), summer (n=4), and autumn (n=4). Progesterone concentrations were summarized by age-gender class as follows: female calves (n=7), female subadults (n=17), non-pregnant female adults (n=30), pregnant female adults (n=15), male calves (n=2), male subadults (n=16), and male adults (n=27). Calves were 26

PAGE 27

defined by a total straight length 245 cm, subadults were 246 to 265 cm, and adults were >265 cm [50]. Laboratory Analysis The high sensitivity and specificity automated chemiluminescent immunoassay analyzer (IMMULITE 1000: Siemens Medical Solutions Diagnostics, Los Angeles, CA) uses a competitive polyclonal rabbit anti-progesterone antibody and two reagents (ligand-labeled synthetic progesterone and alkaline phosphatase conjugated to anti-ligand, in buffer) (product no. LKPG1, Siemens Medical Solutions Diagnostics, Los Angeles, CA). This progesterone assay was analytically validated for manatee serum by testing precision, accuracy, and analytical sensitivity, and by completing a methods comparison experiment. Analytical specificity (assays specificity for progesterone versus interfering compounds) was established by the manufacturer (Siemens Progesterone PILKPG-10, 2006-12-29) but all other parameters were tested at the UF endocrine laboratory. Intra-assay and inter-assay precision were assessed with repeat sample testing and accuracy was assessed with dilution experiments. Assay analytical sensitivity (the lower limit of detection or LLOD) was established during experiments to test precision and accuracy where the assays LLOD was challenged. Results generated by the IMMULITE at UF were compared with a progesterone Coat-A-Count radioimmunoassay (RIA) (Siemens Medical Solutions Diagnostics, Los Angeles, CA) at the Missouri State Universitys Department of Agriculture. The RIA, with a LLOD equal to 0.05 ng/ml, has been validated for use in elephants (D. Schmitt pers. comm.). Samples across the manatees physiologic range of progesterone concentrations were included in validation experiments. Serum samples with a range of progesterone concentrations (n=9) were tested for effects of freeze-thaw upon measured progesterone concentrations in manatee samples. Fresh samples were analyzed within 12 hours of collection by venipuncture. These fresh samples were frozen 27

PAGE 28

to -80C and the same aliquots were thawed twice within one year of their original collection date to assess any changes in measured progesterone concentration. Samples were refrozen immediately following each assay. Statistical Analysis Validation results (precision, accuracy, methods comparison, and ROC) were calculated with EP Evaluator Release 7 (David G. Rhoads Associates, Inc., Kennett Square, PA). Precision was evaluated by coefficients of variation (CV). Accuracy, assessed by linearity, required definition of total allowable error (25%), and linearity was achieved when the dilution results did not differ from the expected results by a percentage greater than the systematic allowable error (50% of the total allowable error budget). Accuracy was further evaluated using adjusted-R 2 values resulting from linear regression, where results closer to 1.0 indicate a better relation between the independent and dependent variables. Methods were compared using Deming regression [56] and results were evaluated based upon a total allowable error (TEa) [57] of 25% and the observed correlation coefficient (r). Diagnostic validation using ROC analysis was completed to determine the assays ability to discern differences between pregnant and non-pregnant manatees [51]. For each ROC analysis, diagnostic sensitivity (tests probability of producing a true positive result) and specificity (tests probability of producing a true negative result), and positive and negative predictive values were determined. Positive predictive value (PPV) is the probability that a positive test result accompanies a true positive condition [# True Positives / (# True Positives + # False Positives)] while negative predictive value (NPV) is the probability that a negative test result accompanies a true negative condition [# True Negatives / (# True Negatives + # False Negatives]). Power calculations (=0.05) were used to calculate the number of samples required to distinguish non-pregnant and pregnant manatees via ROC analysis (Minitab 14, Minitab Inc., 28

PAGE 29

State College, PA). The remaining statistics were completed using SigmaStat (Systat Software, Inc., Point Richmond, CA). A Kruskal-Wallis one-way ANOVA on Ranks was used to investigate differences in progesterone concentration between adult non-pregnant Florida manatees (all with SAA 60 g/ml) that were (n=10) and were not (n=16) lactating (mean SD) to determine if lactating females could be grouped with other non-pregnant females. Additionally, a one-way ANOVA was used to test differences in progesterone concentration between healthy, non-pregnant, non-lactating female manatees from Mexico (n=7) and Florida (n=16) (different subspecies) to determine if females from Mexico could be included in the study population. Paired t-tests were used to test for significant differences in measured progesterone concentration among samples (n=9) that were frozen and thawed. Fresh samples were first compared to samples that had been freeze-thawed once, and a second test compared fresh samples to samples freeze-thawed twice. Dunns method (ANOVA on ranks), which provides pairwise significance testing, was also used to test for variation in progesterone concentrations among calves (n=9), subadults (n=33), and adults (n=58). Statistical tests were deemed significant when P<0.05. Results Assay Analytical and Diagnostic Validation Coefficients of variation (CV) from precision analyses ranged from 6.2% to 14.5%. Accuracy results were linear within the allowable systematic error, with adjusted-R 2 values from 0.93 to 1.0 (Table2-1). The analytical sensitivity of this assay system for manatee serum is 0.1 ng/ml based on dilution curves and precision results. The methods comparison produced a Deming regression slope of 0.95 [95% CI: 0.89 to 1.01] and an intercept of 0.01 [95% CI: -0.06 to 0.09] indicating diagnostically acceptable results (Fig. 2-1). 29

PAGE 30

In the Florida manatee, progesterone concentrations in non-pregnant, lactating females (0.2 0.1 ng/ml; n=10) were slightly higher than those of non-pregnant, non-lactating females (0.1 0.06 ng/ml; n= 16). However, the difference was not statistically significant (P=0.065) and is not believed to be clinically significant. There was also no statistically significant geographic difference (P=0.231) in progesterone concentration between non-pregnant, non-lactating females from Mexico (0.1 0.00; n=7) and Florida (0.1 0.06; n=16). Power calculations indicated that a sample size of 15 non-pregnant and 15 pregnant manatees was required to distinguish pregnancy status via ROC analysis when comparing non-pregnant females to females from throughout gestation. This ROC analysis indicated that assay sensitivity and specificity as well as PPV and NPV are optimized when 0.4 ng/ml is selected as the threshold for pregnancy diagnosis. At the threshold of 0.4 ng/ml, the positive likelihood ratio for pregnancy is 7.0 and this ratio increases to 22.0 at concentrations 0.5 ng/ml (Table 2-2). Power 95% was also achieved within ROC analyses for the first and second halves of gestation. Analysis for the first half of gestation (range: 1.0-5.6 ng/ml) reported sensitivity, specificity, PPV, and NPV as 100% at a threshold of 1.0 ng/ml. For the second half of gestation (range: 0.2 to 1.7 ng/ml), sensitivity was 90.0% and specificity was 86.7% at a threshold of 0.4 ng/ml. The associated PPV was 69.2% and the NPV was 96.3%. Effects of Freeze-Thaw Although there was some variation in progesterone concentration following freeze-thaw, there were no statistically significant differences among the results obtained from fresh samples (2.8 5.2, n=9) and those that were freeze-thawed once (3.0 5.8, n=9, P=0.610) and twice (3.1 5.8, n=9, P = 0.309). 30

PAGE 31

Normal Gestational Changes In a rehabilitating female manatee sampled repeatedly during early pregnancy, serum progesterone concentrations peaked at 4.8 ng/ml in month two of gestation, and declined to 0.9 ng/ml by gestational month seven (Fig. 2-2). For pregnant, free-ranging females, estimation or confirmation of a parturition date allowed a blood sample and its associated progesterone concentration to be linked to the stage of gestation (Table 2-3). Based upon available data, early gestation (months 1 to 4) was characterized by progesterone concentrations ranging from approximately 1.7 to 4.7 ng/ml. During mid gestation (months 5 to 8), progesterone concentrations were near 1.0 ng/ml, and in months 10 and 11 (a portion of late gestation), concentrations between 0.3 and 0.5 ng/ml were measured. Abortion Detection Progesterone data were available for three female manatees that aborted a fetus while being treated for injuries (#1 and #2 = entanglement; #3 = watercraft strike) at rehabilitation facilities (Table 2-4). All females were estimated to be in late pregnancy (the final four gestational months) at the time of their abortion, based on fetal length. Females #1 and #3 displayed baseline serum progesterone concentrations of 0.1 ng/ml prior to abortion. Seasonal Variation Mean progesterone concentrations in pregnant females were six to 20 times greater than those of non-pregnant females for a given season. The highest mean progesterone concentration observed was associated with pregnant females in autumn (3.9 1.8 ng/ml), which was much greater than concentrations observed in pregnant or non-pregnant females during any other season. There was no evidence of seasonal variation in progesterone concentration for non-pregnant females within the sample population (Table 2-5). 31

PAGE 32

Progesterone Concentrations by Gender and Age Class Statistics indicated that pregnant female progesterone concentrations in pregnant adult females (median: 0.8 ng/ml) were significantly higher than concentrations in female and male calves and subadults, non-pregnant adult females, and adult males (Table 2-6). Discussion Assay Analytical and Diagnostic Validation Serum progesterone concentration is a reliable pregnancy diagnostic in species where there are clear differences between progesterone concentrations in diestrus and pregnancy [10,11, 13-15]. Pregnancy in many domestic species is characterized by serum progesterone concentrations much greater than those observed in manatees. For example, the lower threshold for pregnancy diagnosis in the sow is 7.5 ng/ml and concentrations up to 40 ng/ml occur [15]. For this reason, validation of a highly sensitive assay was desired for the manatee. The IMMULITE system, which was developed for human use and validated for several species, including small carnivores (J. Verstegen pers. comm.), was validated for manatees. Any commercially-available progesterone assay must be validated for use in manatees prior to routine use, and assays with a more sensitive LLOD (0.1 ng/ml) are most useful. The progesterone assay was analytically validated, with acceptable coefficients of variation <25% [58], adjusted-R 2 values that yielded linear results for accuracy [59], and a correlation coefficient 0.975 for methods comparison [60] (Table 2-1). For methods comparison, the three results that fell outside of the region of total allowable error occurred below the threshold for pregnancy diagnosis and close to the assays LLOD, where the total allowable error margin is smallest. This error was not clinically significant. In order to create a sample population representative of the free-ranging population, as required for diagnostic validation, 33.3% of samples originated from pregnant females, which 32

PAGE 33

corresponds to the suspected prevalence of pregnancy in the free-ranging manatee population [1]. Progesterone results evaluated by ROC analysis indicated that sensitivity and specificity are both optimized at 0.4 ng/ml, identifying this as the appropriate threshold for diagnosing manatee pregnancy. These values are diagnostic until the final two months of gestation when progesterone concentrations may fall below 0.4 ng/ml. At this time, females are visibly pregnant, making progesterone analysis unnecessary to diagnose pregnancy. Previous attempts to measure progesterone in manatees have involved fecal samples and a limited number of plasma (assay LLOD = 0.08 ng/ml) and serum samples [12,40]. It was found that plasma concentrations were lower than those recovered from fecal samples and there was no correlation between matching fecal and plasma samples [12]. The long gut transit time in manatees and the associated digestion of progesterone result in imprecise and inaccurate fecal hormone concentrations, which complicates the accurate medical assessment of pregnancy using fecal hormones. Effects of Freeze-Thaw Freeze-thaw cycles did not significantly affect progesterone assay results, which is consistent with previous findings [61]. Although some changes were observed throughout the subsequent trials, no clear pattern existed and the differences among the means were not statistically significant. Normal Gestational Changes Though manatee progesterone concentrations overall are lower than those of some other species [15,25], particularly during early pregnancy, manatees can have a five-fold or greater increase in serum progesterone during the first 4 months of pregnancy. We hypothesized that serum progesterone concentrations would decline in the latter stages of pregnancy, and such a decline was observed; concentrations peaked during early pregnancy (1.7 to 4.7 ng/ml) and 33

PAGE 34

became progressively lower during mid (~1 ng/ml) and late (0.3 to 0.5 ng/ml) gestation. Progesterone concentration varies during gestation in other species, such as the mare, where serum progesterone declines to 1-2 ng/ml by days 180-200 of the 340 day pregnancy as a result of conversion to feto-placental 5-reduced progestins, which appear in the circulation by approximately gestational day 70 [25]. In African and Asian elephants, progesterone concentrations are low throughout gestation (0.07 to 1.3 ng/ml) because progesterone metabolites are predominant [17,43,62]. Lower progesterone concentrations observed closer to term in the manatee should not affect the diagnostic utility of this assay, which is most needed to detect pregnancy before visual indicators appear after month six of the 12 month gestation. Until diestrus progesterone concentrations can be fully documented in the manatee, serial sampling of progesterone is suggested to confirm pregnancy, as has been recommended for elephants [44]. If a progesterone result 0.4 ng/ml is measured from a female manatee, a second sample should be submitted for analysis two weeks after the initial sample was collected (based on fecal estimates of manatee estrous cycle duration [12] and gut transit time [42]). If this second sample also measures a progesterone concentration 0.4 ng/ml, the female can be considered pregnant. If the second sample results in a progesterone concentration <0.4 ng/ml, then the original sample likely represented diestrus within that female. This serial sampling technique does not apply to long-term captive females without male contact. Abortion Detection Progesterone concentrations at or near baseline (0.1 ng/ml) appear to be indicative of abortion in the manatee, which is consistent with our hypothesis. In the pregnant mare, abortion is associated with a rapid decline in progestins or progestin concentrations <2.0 ng/ml for more than 3 to 4 days [63]. This threshold is lower in manatees, as evidenced by baseline progesterone concentrations (0.1 ng/ml) observed at least 2 weeks prior to abortion in two late 34

PAGE 35

pregnancy females. Because progesterone concentrations associated with a healthy pregnancy in the manatee also decline to 0.3 to 0.4 ng/ml in the final months of gestation, diagnosing an impending abortion may be more difficult during this time than at an earlier stage of gestation, when progesterone concentrations are higher. For clinicians who are rehabilitating pregnant manatees with such traumatic injuries as watercraft strikes and entanglements, progesterone should be evaluated regularly due to the effects of stress on pregnancy and the possibility of abortion [39,64]. Additionally, in cases where pregnancy is suspected or confirmed, treatment strategies for rehabilitating females should be modified to prevent injury to the fetus by avoiding situations and activities (e.g. removal from the water and lateral rolling) that may result in additional pressure to the abdomen or either lateral aspect of the body. Seasonal Variation Manatees are believed to be diffusely seasonal breeders [1]. The breeding season for the Florida manatee has been proposed to span from March through September [1,24,65] based on the presence of follicles and corpora lutea in the ovaries of female carcasses [24], maximum perinatal carcass recovery [66], and the occurrence of mating herds [1,65]. Patterns in male sperm production also imply decreased likelihood of successful mating during the winter months [67]. In this study, the maximum progesterone concentrations during pregnancy were observed in autumn (September 23 to December 21), which corresponds to mating and implantation between early summer and autumn (based on peak progesterone concentrations observed during early pregnancy in the current study). Descriptive summary statistics are presented without statistical analysis of seasonal variation due to the small sample size (n=1-4) available for most seasons. The observed autumnal progesterone peak for pregnant females was later than expected, compared to previous studies of manatee mating and behavior [1,24,65], and may be 35

PAGE 36

due to this low sample size. The deviation from the expected summer progesterone peak was caused in part by two females in late pregnancy (with low progesterone concentrations) in the summer population. Because the mating and birthing seasons overlap [24] and because serum progesterone concentrations in the manatee decline throughout gestation, seasonal values are affected by the presence of females at different gestational stages, with variable progesterone concentrations. As hypothesized, the highest progesterone concentrations were measured in pregnant females during all seasons of the year. We also hypothesized that seasonal differences in progesterone concentration would be apparent in the non-pregnant population, with higher concentrations associated with episodes of diestrus during the breeding season. However, no appreciable seasonal differences existed in the non-pregnant population, for which the mean progesterone concentrations were at baseline (0.1 ng/ml) for all seasons except autumn, where the mean was 0.2 0.1 ng/ml. The observed results suggest that the non-pregnant sample population contained only a small number of diestrus females, which eliminated the possibility of viewing seasonal progesterone differences within the non-pregnant population. Progesterone Concentrations by Gender and Age Class It was determined that pregnant adult females exhibit the highest progesterone concentrations in the manatee population, as expected. These findings concur with a previous study of manatee progesterone [12]. Future Studies Additional studies of manatee progesterone and pregnancy should monitor concentrations of this hormone throughout gestation, which would require sample collection from well-trained, healthy, captive or rehabilitated females, to minimize stress associated with handling. Analysis of serum progesterone throughout non-pregnant estrous cycles is also required to determine the 36

PAGE 37

duration and range of progesterone concentrations that characterize non-pregnant diestrus. Such an investigation would require the use of captive females housed with males to help guarantee normal reproductive cyclicity. Conclusion The IMMULITE 1000s progesterone assay was determined to be a valid and highly sensitive and specific tool for diagnosing pregnancy in the manatee, particularly during early and mid gestation. Progesterone analyses following one and two freeze-thaw cycles did not significantly affect assay results. Changes in serum progesterone concentrations throughout manatee gestation were identified, with the highest concentrations occurring during early pregnancy. The progesterone concentration associated with abortion in two late pregnancy females was also identified (0.1 ng/ml). The highest progesterone concentrations in pregnant females were observed in autumn and pregnant adult females exhibited the highest progesterone concentrations in the manatee population. This assay will allow manatee biologists and veterinarians to monitor reproductive health in individual free-ranging and captive females and will be beneficial in managing captive breeding of Sirenians. Table 2-1. Manatee Serum Progesterone Analytical Validation Results Validation Test Range of Sample Concentration Means Tested SD (ng/ml) (n) Number of Replicates Validation Result Inter-assay Precision 0.9 0.2 to 17.2 0.2 (6) 3 CV min-max: 6.2 to 14.5% Mean CV = 9.2% Intra-assay Precision 1.2 0.1 to 7.2 0.5 (3) 13 CV min-max: 7.2 to 10.9% Mean CV = 9.7% Accuracy 0.3 to 4.6 (3) 5 (dilutions) Adjusted-R 2 min-max: 0.93 to 1.0 Mean Adjusted-R 2 : 0.98 Methods Comparison 0.1 to 5.8 (40) 1 r = 0.98 37

PAGE 38

Table 2-2. Progesterone (P4) Concentrations for Pregnancy Diagnosis throughout Gestation P4 Cutoff (ng/ml) Sensitivity Specificity Positive Predictive Value (PPV) Negative Predictive Value (NPV) Positive Likelihood Ratio 0.1 100.0% 0.0% 33.3% -1 1.0 0.4 93.3% 86.7% 77.8% 96.3% 7.0 0.5 73.3% 96.7% 91.7% 87.9% 22.0 1.0 46.7% 100.0% 100.0% 78.9% -2 1 There is no NPV at 0.1 ng/ml because this threshold encompasses all progesterone results. 2 The Positive Likelihood Ratio is calculated as sensitivity / (1-specificity), which results in a denominator of zero at 1.0 ng/ml. Table 2-3. Gestational Progesterone Concentrations Determined from Study Females Gestation Month Average Progesterone SD (ng/ml) Number of Manatees Sample Size 1 1.7 1 1 2 3.6 1.3 1 4 3 4.7 1.2 3 3 4 1.7 1 1 5 0.8 1 1 6 1.0 1 1 7 0.8 0.1 4 4 8 1.3 0.5 2 2 9 No Data --10 0.4 0.06 3 3 11 0.3 0.07 1 2 12 No Data --Table 2-4. Abortion-Associated Serum Progesterone (P4) (ng/ml) in Late Pregnant Females Sampling Time Frame Female #1 Female #2 Female #3 At Rescue 0.1 0.3 0.1 Within 2 Weeks of Rescue 0.1 0.4 0.1 Post-Abortion 0.1 (n=1) -0.1 (n=4) Normal Late Pregnancy 0.3-0.4 0.3-0.4 0.3-0.4 Fetal Length at Abortion (cm) 109 115 124 Table 2-5. Seasonal Progesterone Values (mean P4 SD) for Non-Pregnant and Pregnant Florida Manatees Season Non-Pregnant (ng/ml) (n) Pregnant (ng/ml) (n) Winter 0.1 0.0 ( 3) 0.7 0.2 (4) Spring 0.1 (1) 0.6 0.3 (2) Summer 0.1 0.09 (5) 0.6 0.3 (4) Autumn 0.2 0.1 (12) 3.9 1.8 (4) 38

PAGE 39

Table 2-6. Progesterone Concentrations by Gender and Age Class Gender and Age Categories Mean SD (ng/ml) (n) Median (ng/ml) Min-Max (ng/ml) Female Calf 0.3 0.3 (7) b 0.1 0.1 0.9 Female Subadult 0.1 0.1 (17) b 0.1 0.1 0.7 Non-Pregnant Female Adult 0.2 0.2 (30) b 0.1 0.1 0.9 Pregnant Female Adult 1.2 1.3 (15) a 0.8 0.2 5.3 Male Calf 0.1 0.1 (2) b 0.1 0.1 0.2 Male Subadult 0.2 0.2 (16) b 0.1 0.1 0.7 Male Adult 0.2 0.1 (27) b 0.1 0.1 0.6 a Pregnant female adults had significantly higher progesterone concentrations than any other age class of females or males b The assay lower limit of detection (LLOD) is 0.1 ng/ml. 6543210 UF IMMULITE (ng/ml)6 5 4 3 2 1 0 TEa Elephant Validated RIA (ng/ml) Figure 2-1. Scatter plot for methods comparison. The plot shows progesterone results for paired samples tested with both a radioimmunoassay (RIA) validated for use in elephants (x-axis) and the chemiluminescent system (y-axis). The progesterone results are represented by the points on the graph and the yellow area of the plot represents the total allowable error (TEa) for the methods comparison, which was set at 25%. Points outside the yellow region (red) represent results with error that exceeds the defined TEa. A perfect fit between the methods is represented by results that fall along the line in the center of the yellow region. 39

PAGE 40

Estimated Time of Ges tation 10/1/99 1/00 1/00 1/00 2/6/10/ Progesterone (ng/ml) 0123456 Progesterone Values (ng/ml) Through Pregnancy Parturition 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Estimated Month of Gestation Figure 2-2. Progesterone profile generated from a single female of known gestation. Blood sampling was routine until the fourth month of gestation, after which only one additional sample was collected prior to parturition. The arrow indicates the date of parturition. During the first two months of sampling (encompassing gestational months one through three), progesterone concentrations fluctuated, ranging from 1.2 to 4.8 ng/ml. Samples collected during months four and six had measured progesterone concentrations of approximately 1.0 ng/ml. A final sample was collected 50 days after parturition and a progesterone concentration of 0.1 ng/ml was measured. 40

PAGE 41

CHAPTER 3 DETECTION AND IDENTIFICATION OF PROGESTERONE METABOLITES IN THE FLORIDA MANATEE USING GC/MS/MS Introduction Progesterone (P4) is a steroid hormone with the typical four-ring structure of pregnenolone. Progesterone synthesis and metabolism are associated with changes to the accessory groups of the pentagonal D ring, producing progesterones various metabolites. Although the chemical differences among these steroids and metabolites are subtle, they have significant implications when evaluating luteal function and measuring circulating progestin concentrations. Furthermore, the spectrum of steroids produced by the corpus luteum may vary dramatically among species. In African and Asian elephants, which are close terrestrial relatives of the manatee [7,8], progesterone (pregn-4-ene-3,20-dione; P4) concentrations during pregnancy peak below 2.0 ng/ml [16]. These observed concentrations are significantly lower than those of many species [10,13-15]. This may be caused by varied steroidogenic pathways and progestins required for pregnancy maintenance or rapid metabolism of progesterone, as occurs in carnivores [68]. Using gas chromatography-mass spectrometry (GC/MS), different progesterone metabolites have been detected in African and Asian elephants, including the 5-reduced progestins: 5-pregnane-3,20-dione (5 -DHP) and 3-hydroxy-5-pregnan-20-one (5-P3-OH) [17]. In the Asian elephant, 17-hydroxyprogesterone (17-OHP) was also detected [18]. In both species of elephant, 5 -DHP predominates, with concentrations of up to 21 ng/ml observed during pregnancy [69]. In the mare, 5-DHP and other 5-reduced progestins predominate during the second half of gestation, while P4 concentrations are minimal [19]. Finally, 20-OHP (20-hydroxy-4-pregnen-3-one; 20-hydroxyprogesterone), an inactive form of P4, has been observed in species including the rat and bovine [70,71]. In the rat, circulating concentrations of 41

PAGE 42

20-OHP are relatively stable throughout all stages of the estrous cycle and pregnancy [70], while concentrations increase during the latter stages of bovine pregnancy [71] (Fig. 3-1). Circulating P4 concentrations observed in the manatee (Trichechus manatus) are similar to those of elephants, but may be higher during the first four months of pregnancy (see Chapter 2). These manatee P4 concentrations may indicate rapid production and metabolism of progesterone or low levels of production coupled with a slow rate of metabolism, resulting in prolonged endogenous availability of a low level of progesterone. The possibility of an alternate pattern of steroidogenesis leading to the production of biologically active metabolites, as occurs in elephants, cannot be excluded. Finally, a paracrine secretion a nd transport of progestins from the ovary to the uterus may occur. This would result in increased concentrations of progesterone localized in the reproductive tract that may not be measured in peripheral blood samples. The objectives of this study were to use gas chromatography with tandem quadrupole detectors (GC/MS/MS) to: (1) determine if the relatively low P4 concentrations observed in the manatee are associated with a high ratio of 5-pregnane-3,20-dione (5 DHP), 3-hydroxy-5-pregnan-20-one (5-P3-OH), 17-hydroxyprogesterone (17-OHP), or 20-hydroxy-4-pregnen-3-one (20-OHP); and (2) document any changes in metabolite presence or relative concentration associated with gender or the various stages of pregnancy or diestrus. Materials and Methods Blood Sampling Manatee blood samples were collected during health assessments of Florida manatees from the years 2001 through 2007. Samples were opportunistically collected during ongoing ecological investigations [46,47] and from captive manatees housed at facilities throughout Florida. Lithium heparin plasma and serum were separated within one hour of collection by onsite centrifugation (3,000 x g for 10 minutes), then transported on ice to the University of 42

PAGE 43

Florida. Lithium heparin plasma was stored at -80C until shipment for GC/MS/MS analysis, and serum was either assayed immediately for P4 or stored at -80C until analysis. Sample Population Lithium heparin plasma samples from 24 adult [50] Florida manatees were selected for GC/MS/MS analysis. The sample population included free-ranging (n=3) and rehabilitating (n=3) males; non-pregnant and non-lactating captive (n=5) and free-ranging (n=1) females; lactating free-ranging females (n=4); and early (n=2), mid (n=3), and late pregnant (n=2) free-ranging females. One non-pregnant, non-lactating, severely injured, rehabilitating female diagnosed with polycystic ovaries was also included. The condition of this females ovaries was observed at necropsy <2 hours after blood collection. Four of five captive females in the study population were believed to be reproductively abnormal due to known health problems and the absence of male interaction and mating in a female-only captive environment. Pregnant females were identified by serum P4 concentration obtained from immunoassay as well as subsequent field observation of a calf within one year of capture. The stage of gestation (early, mid, or late) was determined based on knowledge of a 12 month gestation [1], observations that calves may be at lengths from 82 to 160 cm or 95 to 155 cm [1,24], and an assumption that fetal growth is approximately linear. Progestins observed in captive females were considered to be associated with physiologic or pathologic luteal development. Laboratory Analysis Manatee serum was first assayed to determine its P4 concentration using an automated chemiluminescent assay system analytically validated for use in manatees (see Chapter 2). Manatee plasmas were sent overnight in a dry shipper to the CER Endocrine Laboratory (Belgium) for GC/MS/MS analysis as previously described for other steroids [72]. Briefly, a standard analysis procedure including liquid-liquid extraction of plasma samples combined with 43

PAGE 44

GC/MS/MS analysis was performed using a HP6890 gas chromatograph (Agilent Technologies, Brussels, Belgium) coupled with a Quattro Micro GC mass selective detector (Micromass, Altricham, Cheshire, UK). The Quattro Micro is a benchtop GC/MS/MS system that provides low-level quantitative analysis and limited matrix interference [73]. Samples as well as standards to confirm the presence of particular compounds were injected after extraction pre-treatment. The standards used were P4, 5-pregnane-3,20-dione (5 DHP), 3-hydroxy-5-pregnan-20-one (5-P3-OH), 17-hydroxyprogesterone (17-OHP), and 20-hydroxy-4-pregnen-3-one (20-OHP) (Sigma-Aldrich Company, Steinheim, Germany). The GC column was an HP5-MS (5% phenyl, methyl silicone; product number 190915-433) bonded phase fused silica capillary column (30 m x 0.25 mm i.d., 0.25 m film thickness) (Agilent Technologies, Brussels, Belgium). The injector temperature was 300C and the injection volume was 1 l. Injections were made in the splitless mode. The temperature inside the gas chromatograph (oven temperature) was initially maintained at 100C for one minute, and was programmed to rise at a rate of 11.5C per minute until reaching a temperature of 325C. The temperature was maintained at 325C for five minutes. The transfer line temperature was set at 320C. The carrier gas, N60 helium, was used at a constant pressure of 100 kPA. The various progestagen steroids were identified in the MRM (multiple reaction monitoring) mode. The source pressure was approximately 2x10 5 Pa and the collision pressure at least 2.5x10 3 Pa to achieve the highest sensitivity. Data were collected using a HP Chemstation G1701BAB01.00. Evaluation of the steroids was performed by a liquid-liquid extraction using 1ml of plasma, 1ml of water, and 50l of Td3 (16, 16, 17-d 3 -testosterone; an internal standard). The extraction products (2 ml) were then added to the Extrelut NT3 (Merck, Darmstadt, Germany) column for 5-10 minutes, and eluted with 3x5 ml ethyl acetate. The organic phase was evaporated at 40C 44

PAGE 45

and recovered with 100 l methanol (MeOH). Three ml of water were added to the product. A clean up procedure was completed by conditioning the column (C18, 500 mg, 3 cc tube) with 5 ml MeOH and passing through 5 ml water. The extracted sample was added to the column, which was then rinsed with 2 ml water. Next, the column was washed with 5 ml water, 5 ml of a MeOH and water mixture (20/80), and 2.5 ml hexane. The column was dried and eluted with 2x1.5 ml of ethyl acetate. The ethyl acetate was evaporated at 40C and derivatized with 25 l of a stock standard solution of derivatization agent prepared as follows: 2 mg of dithiothreitol were mixed with 1 ml of MSTFA (N-Methyl-N-trifluoroacetamide) and 17 l of trimethylsilyliodotrimethylsilane. This reagent gives rise to trimethylsilylenol and trimethylsilyl ether derivatives. After vortex mixing, the mixture was heated at 60C for 30 minutes and 2 l of product were injected into the gas chromatograph column. Each sample was injected twice and the profile analyzed to determine the presence of the various progestins. Data Analysis The use of GC/MS/MS provides accurate identification of compounds, but the concentrations determined in this study should be used for intra-assay data comparison only. These values are not comparable to progesterone immunoassay results (Chapter 2). Within the text, progestins measured with GC/MS/MS are described as not detected (ND), at the lower limit of detection (LLOD), trace (2x LLOD), or low, moderate, or high (relative to the total measured concentrations). Results Identification of the Predominant Progestin Among the 24 plasma samples tested, P4 was observed most frequently (n=14) (Table 3-1). The second most frequently observed form of progesterone was 5-P3-OH (n=12), followed by 5DHP (n=7). The 17-OHP metabolite was observed in only three samples, 45

PAGE 46

including two males and one female with numerous cystic ovarian follicles and corpora lutea. One sample exhibited 20-OHP at a concentration equal to the LLOD for that metabolite (0.01 ng/ml): a non-pregnant, non-lactating captive female (Table 3-1). Among non-pregnant, non-lactating females, P4 comprised 57.0% of the observed progestins, followed by 5-P3-OH (31.6%) and 17-OHP (10.9%). The progestins 5DHP and 20-OHP comprised only 2.1% and 0.05% of the observed progestins in non-pregnant, non-lactating females, respectively. Among pregnant females, P4 and 5-P3-OH comprised similar percentages of total percentages as in non-pregnant, non-lactating females (55.6% and 31.1%, respectively) and also exhibited 5DHP (13.2%). The only additional peaks observed during chromatographic analysis represented isomers of the standards. Gender and Reproductive Differences in Progestins For the six males tested, P4 was detected in one male at LLOD, and 17-OHP was observed in three males, ranging from LLOD to trace. Among five captive, non-pregnant, non-lactating females, P4 (n=3; LLOD to high), 5-P3-OH (n=5; trace to high), 5DHP (n=2; low to moderate), and 20-OHP (n=1; LLOD) were detected. For one free-ranging, non-pregnant and non-lactating female, low concentrations of P4 and trace concentrations of 5-P3-OH were observed. A severely injured, rehabilitating female with cystic follicles and corpora lutea exhibited moderate P4, 5-P3-OH, and 5DHP and LLOD 17-OHP. In three out of four non-pregnant, lactating females, no P4 or metabolites were detected, while one of these lactating females possessed a LLOD concentration of P4. Two early pregnant females possessed moderate to high P4, low to high 5-P3-OH, and LLOD to moderate 5DHP. Among three mid pregnancy females, low P4, and LLOD 5-P3-OH and 5DHP were observed. The two females in late pregnancy exhibited low P4 and one measured a LLOD concentration of 5-P3-OH. In 20 of the 24 manatees studied, metabolites were not observed if P4 was not 46

PAGE 47

observed. A similar relationship occurred for 5-P3-OH and 5DHP, where only one female (in mid pregnancy) exhibited 5DHP in the absence of 5-P3-OH. The most frequent occurrences of P4, 5-P3-OH, and 5DHP were associated with early pregnancy and physiologic or pathologic luteal development in captive females (Table 3-1). Discussion Identification of the Predominant Progestin Manatee plasma analyzed by GC/MS/MS resulted in identification of several P4 metabolites, but P4 was the predominant progestin. The GC/MS/MS analysis did not detect any additional progestins other than those for which standards were included. We hypothesized that P4 metabolites in the manatee could be identified using GC/MS/MS, which we confirmed. However, we also hypothesized that metabolites contributed significantly to total progestin content in the manatee. Contrary to our hypothesis, the lower relative concentrations of circulating metabolites do not suggest that rapid metabolism or an alternate steroidogenic pathway is responsible for the observed progesterone concentrations. Metabolites present in manatee plasma were identified, and were similar to those of elephants and hyraxes, although their role in the manatee is likely less significant than within these other related species [7,8] based on the minimal relative concentration of metabolites observed in manatees. In hyrax plasma, for example, 5-DHP was measured at concentrations ranging from 3.5 to 9.6 ng/ml, which were 4 to 17 times greater than measured P4 concentrations (0.4 to 5.9 ng/ml) [74]. In manatee plasma, 5-DHP was measured at concentrations ranging from LLOD to moderate, which were up to six times less than measured P4 concentrations. Findings in peripheral circulation of the manatee do not suggest that progestin metabolites are biologically active in this species. However, manatees may utilize paracrine transport of progestins, preventing the true concentration of progestins from being measured in peripheral 47

PAGE 48

circulation. Male manatees possess vasculature believed to cool the intra-abdominal epididymides via counter-current heat exchange to prevent hyperthermic insult. This vasculature consists of an arteriovenous iliac vascular bundle and an inguinal venous plexus, and similar vasculature has also been observed in female manatees [75]. While counter-current vascular exchange often involves the transfer of heat, substances will also diffuse from veins to arteries down concentration gradients [76]. Steroid hormones such as P4 have been determined to be capable of counter-current exchange due to their size (250-300 Da) and lipophilic nature [77], and it may be possible that manatees utilize an enhanced counter current exchange mechanism to deliver progestins from the ovaries to the uterus and placenta (paracrine tr ansport). If manatee progestins were transported through such a localized vascular system, this could explain the relatively low concentrations observed in the peripheral circulation. We hypothesized that manatees, like elephants, would possess 5-P3-OH, 5DHP, and 17-OHP in peripheral circulation. In equines, elephants, and hyraxes, 5DHP is believed to be a biologically active metabolite due to its high binding affinity for that species endometrial progestin receptors [74,78]. In the horse, this receptor affinity is believed to be greater than that of P4 itself [79]. The 5DHP metabolite is also the most predominant circulating progestin in elephants [16,17]. However, 5-P3-OH, which is predominant in elephant luteal tissue [17], is considered to be an inactive metabolite [16]. In the manatee, 5-P3-OH was the most frequently observed metabolite, while 5DHP was observed in concentrations that ranged from LLOD to moderate, suggesting a less significant role in the manatee, as compared to the elephant. With less than half the biological activity of P4 [80], 20-OHP is considered to be a mechanism to reduce progesterones influence on reproductive tissues [70]. In the diestrus rat, the concentration of 20-OHP is 4-6 fold greater than P4. Bovine cotyledons also contain 20-OHP, 48

PAGE 49

which increases and peaks in late gestation (months 7 to 8) and subsides after parturition (month 10) [71]. We hypothesized that relatively low circulating P4 concentrations could be associated in part with increased 20-OHP concentration. However, 20-OHP was only measured in one manatee (at LLOD), refuting the hypothesis that low circulating progesterone concentration could be explained by over-production of 20-OHP leading to progesterone inactivation. It has been determined that manatees do not have high circulating concentrations of metabolites, and serum P4 concentration during pregnancy rarely exceed 5.0 ng/ml (see Chapter 2). If previously measured serum P4 concentrations represent the amount of steroid responsible for maintaining pregnancy in this species, this may suggest that manatees use the P4 they produce very efficiently, leaving only minimal amounts of progestins free in the peripheral circulation. Additionally, the near-absence of metabolites and decline in serum progesterone concentration during the second half of gestation (Chapter 2) suggests that the manatee placenta does not contribute to progestin production that supports pregnancy. However, the production of a chorionic gonadotropin or other non-progestin based factor by the placenta may be possible. The guinea pig, like the manatee, exhibits a significant increase in progesterone concentration during the beginning of pregnancy. In the guinea pig, this increase is accompanied by a significant decline in metabolic clearance rate and an increased concentration of specific binding proteins in blood, which protect progesterone from metabolism [81]. This mechanism to conserve progesterone has also been demonstrated in an herbivorous, semi-aquatic rodent, the coypu or nutria (Myocastor coypus). In the coypu, an increased concentration of progesterone is observed in the blood, but is not necessarily accompanied by an increased secretion rate [82]. While the low relative concentrations of metabolites in manatee plasma suggest the use of mechanisms to reduce metabolism, low concentrations of progesterone itself suggest that the 49

PAGE 50

manatee utilizes some additional mechanism to support pregnancy, such as paracrine progestin transport or a chorionic gonadotropin. A previous study (Chapter 2) determined manatee serum progesterone (P4) concentrations. The assay used, which is highly specific for progesterone, does report a minor cross reactivity with 17-OHP (1.04%) (Siemens Progesterone PILKPG-10, 2006-12-29). The cross reactivities of the other metabolites investigated in the current study have not been tested, and may contribute to measured P4 concentration due to mnor cross reactivity, but this is not likely to be clinically significant. Gender and Reproductive Differences in Progestins There were differences in the metabolites observed among manatee genders and reproductive stages. We hypothesized that the presence and abundance of metabolites would differ with reproductive status, which the results confirmed. The absence of progestins in three of four lactating females, two of which had relatively small calves (177 cm and 183 cm in total straight length, respectively) [1,24,50], suggests that lactational anestrus may occur in the manatee. It is known that lactation and estrus may occur simultaneously in the dugong [83], but it appears that the manatee may experience anestrus, at least during early lactation. In elephants, the 5-reduced progestins are also dominant during diestrus, indicating that these are the dominant luteal secretory product at all times [17]. In non-pregnant manatees, however, P4 was typically more prevalent than 5-P3-OH and 5-DHP. In a female diagnosed with polycystic ovaries, progesterone and metabolite concentrations were lower than concentrations in two non-pregnant, non-lactating captive females, and may indicate the presence of polycystic ovaries or another reproductive abnormality in these two captive females. Observed P4 and metabolite concentrations in these captive females may also represent the normal range of these hormones during diestrus. The relative concentrations of observed 50

PAGE 51

progestins in a free-ranging, non-pregnant and non-lacting female were much lower than those observed in two of the captive females. Because the manatee estrous cycle has not been characterized using serum or plasma progestins, the true implications of these results are unknown. Among pregnant manatees, P4 and 5-P3-OH were most prevalent, although 5-DHP was also present in several females. Metabolites and P4 declined throughout pregnancy so that only very low concentrations of 5-reduced progestins were detectable by mid and late gestation. Plasma concentrations of the 5-reduced progestins in African elephants were also lower in late pregnancy than earlier in gestation, suggesting a decline in production and secretion of these metabolites from the corpus luteum [17]. However, in elephants, plasma concentrations of P4 did not change throughout gestation [17], while P4 concentrations in the manatee decreased (Chapter 2). Zebras and other equines have a different pattern of metabolite expression than manatees and elephants, with 5-reduced metabolites of P4 becoming dominant in the circulation during the second half of gestation [84]. The detection and identification of P4 metabolites in elephants indicated that these species did not lack steroidogenic function, but used a different progestin to support pregnancy [17] and that overall circulating steroid concentrations were really no lower than those of other species [16]. The observed results for the manatee suggest that overall circulating steroid concentrations may be lower than those of other species because metabolites, while present, were even less prevalent than P4 itself. However, the possible contribution of paracr ine circulation warrants further study. Future Studies This study demonstrated that P4 metabolism occurs in the manatee, but in a different pattern than what has been observed in other species [17,69-71,74,79]. It would be beneficial to 51

PAGE 52

determine if the progesterone increase associated with early pregnancy in the manatee is the result of a decreased metabolic clearance rate of progesterone, as occurs in the guinea pig [81]. Additionally, observed circulating P4 concentration in the manatee may be related to this species reproductive and vascular anatomy combined with its unique physiology. Therefore, the vasculature of freshly dead female manatees, particularly pregnant females, should be examined as a possible portal for steroid hormone transfer from the ovaries to the uterus and placenta. Progesterone analysis of tissues collected from a variety of sites in this region could further confirm the route of steroid hormone transfer in the manatee. Finally, an investigation of chorionic gonadotropin in the manatee placenta may help explain the circulating progestin concentrations observed in this species. Conclusion The relatively low circulating P4 concentrations in the manatee cannot be attributed to metabolism, because GC/MS/MS found that P4 itself was predominant in the plasma. The two most commonly observed metabolites were 5-P3-OH and 5-DHP. These metabolites were most prevalent during physiologic or pathologic luteal development in non-pregnant females and early pregnancy, concurrent with maximum P4 concentrations. These results, based on peripheral circulation, suggest a limited role for progesterone metabolites in the manatee. 52

PAGE 53

Table 3-1. Manatee GC/MS/MS Results for Progesterone and Its Metabolites ID Origin C=captive F=free R=rehab. Gender/Status P=pregnant NP=non-pregnant L=lactating NL=non-lactating P4 LLOD: 0.07 ng/ml 5-P3OH LLOD: 0.1 ng/ml 5DHP LLOD: 0.2 ng/ml 17OHP LLOD: 0.1 ng/ml 20-OHP LLOD: 0.01 ng/ml 1 R Male ND ND ND 0.1 ND 2 R Male ND ND ND 0.2 ND 3 R Male ND ND ND ND ND 4 F Male ND ND ND ND ND 5 F Male ND ND ND ND ND 6 F Male 0.07 ND ND ND ND 7 C* NP, NL ND 0.4 ND ND ND 8 C* NP, NL 0.2 0.2 ND ND ND 9 C* NP, NL 6.1 2.6 1.0 ND ND 10 C* NP, NL ND 0.2 ND ND ND 11 C NP, NL 3.2 1.4 0.5 ND 0.01 12 F NP, NL 0.3 0.2 ND ND ND 13 R Cystic, NP, NL 1.2 1.1 0.6 0.1 ND 14 F NP, L ND ND ND ND ND 15 F NP, L ND ND ND ND ND 16 F NP, L ND ND ND ND ND 17 F NP, L 0.07 ND ND ND ND 18 F Early P 0.8 0.5 0.2 ND ND 19 F Early P 5.8 3.9 1.4 ND ND 20 F Mid P 0.4 0.1 ND ND ND 21 F Mid P 0.5 0.1 0.2 ND ND 22 F Mid P 0.4 ND 0.2 ND ND 23 F Late P 0.3 0.1 ND ND ND 24 F Late P 0.2 ND ND ND ND Concentrations are not comparable to immunoassay results, but intra-assay comparisons may be used to determine relative progestin concentrations. Captive females hypothesized to be reproductively abnormal are denoted by *. LLOD=systems lower limit of detection. ND= no metabolite detected. 53

PAGE 54

pregn-4-ene-3,20-dione (progesterone; P4) 3-hydroxy-5-pregnan-20-one (5-P3-OH) 5-pregnane-3,20-dione (5 -DHP) 17-hydroxyprogesterone (17-OHP) 20-hydroxy-4-pregnen-3-one (20-hydroxyprogesterone; 20-OHP) Figure 3-1. The molecular structures of progestins investigated in manatee plasma. 54

PAGE 55

CHAPTER 4 MORPHOLOGICAL CHARACTERIZATION OF CORPORA LUTEA AND LUTEAL CELLS IN THE FLORIDA MANATEE Introduction The mammalian corpus luteum (CL) develops during the diestrus portion of the estrous cycle, and typically arises from an ovulated follicle. After ovulation, the remaining cells proliferate and experience morphological changes to form a CL [20], which consists of lutein cells surrounded by a matrix of connective tissue that includes fibroblasts and numerous capillaries lined with endothelial cells [85]. If an estrous cycle culminates in the fertilization and implantation of an ovum, the CL must remain fully functional to support early pregnancy. However, in the case of an infertile estrous cycle, the CL regresses quickly, as in the cow, where the 8-12 g CL is fully reabsorbed within 4 to 6 days of the start of regression [86]. As a monotocous species, the cow produces only one CL during the estrous cycle and pregnancy [23]. However, there are other monotocous domestic species, such as the horse, which form accessory corpora lutea to aid in the maintenance of pregnancy [25]. Corpora lutea contain at least two types of steroidogenic cells [21] that are derived from granulosa and theca cells of the pre-ovulatory follicle, and are often distinguished by size [87-89]. Among mammals, granulosa lutein cells (GLCs), or large luteal cells, are typically near 25 m in length, with a central nucleus and at least one nucleolus. The cell and nuclear shapes are roughly circular and the cytoplasm stains lightly [90]. These GLCs are characterized by large amounts of smooth endoplasmic reticulum and clusters of mitochondria, which indicate these cells as the site of progesterone synthesis [22,23]. Rough endoplasmic reticulum, Golgi apparatus, and secretory granules also occur in GLC cytoplasm. Theca lutein cells (TLCs), or small luteal cells, are smaller and lengthier, with cup-shaped nuclei [23] bordered by heterochromatin, and exhibit darker staining cytoplasm and nuclei than GLCs. The GLCs are 55

PAGE 56

predominant in most species, but extensive numbers of TLCs are present in species including bovines, porcines, and ovines [91]. Near the end of pregnancy or a non-fertile estrous cycle, corpora lutea degenerate via luteolysis [92,93]. This regression is facilitated by various lysosomes and macrophages that break down or remove degenerating cellular organelles [90,94-96]. In the ewe, the first of these degenerative changes is accompanied by a reduction in luteal cell secretory activity [97]. Several days after the start of luteolysis, the luteal cells themselves decrease significantly in size, lose their uniform shape, and become fragmented [85,98]. In bovines, degenerating mitochondria develop electron dense regions that give the organelles a black color when viewed with electron microscopy [99,100]. Osmiophilic lipid droplets also increase in number within the cytoplasm during this time [85,98,100], representing the continued ability of cells to manufacture lipid, but an inability to alter it into steroid hormones [101]. In mares, highly dilated nuclear envelopes and large quantities of heterochromatin also characterize degeneration [95]. Additionally, the vascular system of numerous species exhibits degenerative changes including obstruction of capillary lumens by distended endothelial cells [95,97,102,103] to reduce blood flow through the CL and facilitate regression [95]. Finally, connective tissue elements such as fibroblasts and collagen become prominent in the intercellular space [95]. Some gross and histological descriptions of manatee corpora lutea have been made [104]. First, it is known that diestrus and pregnancy are characterized by multiple but variable numbers of luteal structures on one or both ovaries, and that multiple ovulations are common in a single estrous cycle, even though the manatee is typically monotocous, with only rare occurrences of twinning [1,24]. Also, a previous study identified that two luteal cell types exist in the manatee: primary luteal cells (13-27 m in length) and secondary luteal cells (6-16 m in length) [104]. 56

PAGE 57

However, changes to cellular morphology throughout diestrus or pregnancy have not been documented in the manatee and endothelial cells and fibroblasts have not been described morphologically. Further knowledge of the structure and function of manatee corpora lutea is required to better understand the estrous cycle and pregnancy in this marine mammal. To help achieve this goal, this study sought to: (1) provide a general description of the manatee CL and identify trends related to CL number and distribution, in one or both ovaries; (2) identify and distinguish luteal cell types that comprise the CL; (3) document endothelial cell and fibroblast morphology in corpora lutea; and (4) describe changes to the corpus luteum and its cells at various states of diestrus and pregnancy, including degenerative changes observed ultrastructurally with transmission electron microscopy (TEM). Materials and Methods Sample Collection Florida manatee (Trichechus manatus latirostris) ovarian tissue was obtained from the Florida Fish and Wildlife Conservation Commissions Marine Mammal Pathobiology Laboratory in St. Petersburg, Florida and analyzed at the University of Florida. Ovaries were removed at the time of necropsy, and pregnancy was assessed by locating a fetus in one of the uterine horns. Where a fetus could not be detected, corpora lutea observed on the ovaries were said to be associated with a non-pregnant diestrus. Corpora lutea were collected in 10% neutral buffered formalin (NBF) and transferred to phospho-buffered saline (PBS) after 24 hours. Sample Population Corpora lutea from five adult (>265 cm total straight length) and one subadult (246-265 cm) [50] female Florida manatees were used in this study, including two diestrus (#1 and #2) and four pregnant (#3 to #6) females. Four specimens were assessed as fresh at the time 57

PAGE 58

of necropsy (#1, #2, #4, and #5), and two were moderately decomposed (#3 and #6). For specimens obtained from pregnant females, stage of pregnancy was assessed by measuring fetal length: early pregnancy <30 cm (n=0); mid pregnancy 31-80 cm (n=2); late pregnancy >81 cm (n=2). These estimates of pregnancy stage and fetal length were based upon knowledge of the 12 month gestation [1] and observations that calves may be born ranging in length from 82 to 160 cm or 95 to 155 cm [1,24]. Samples for TEM were collected from female #4. Sample Processing Tissues were processed, embedded in paraffin, and cut at 5 m. Sections applied to glass slides were placed in a 60C oven for 1-2 hours, then treated with xylene to remove excess paraffin, rehydrated in a series of alcohols, and stained with hematoxylin and eosin (H&E). Samples for transmission electron microscopy (TEM) were sliced into 3 mm sections, placed in 2% glutaraldehyde and refrigerated for approximately 2 weeks until osmication. To begin osmication, tissues were placed in glass vials containing 0.1 Molar (M) sodium cacodylate buffer for two, 15 minute changes. Vials were capped at all times. Next, tissues were placed in a 1:1 mixture of 0.1M sodium cacodylate and osmium tetroxide for 1 hour. After osmication, tissues were placed in two additional, 15 minute changes of 0.1M sodium cacodylate buffer. Tissues were then dehydrated with increasing concentrations of ethyl alcohol (EtOH) (15 minutes each in 25%, 50%, 70%, 2x 95%, and 2x 100%). To begin the embedding process, tissues were placed in mixtures of EtOH and acetone for 15 minutes each (2:1, 1:1, and 1:2), followed by two, 15 minute changes in 100% acetone. Next, tissues were added to a 2:1 mixture of acetone and epon-araldite (araldite 502 / embed-812 embedding media, Electron Microscopy Sciences, Hatfield, PA) for 30 minutes, followed by placement in a 1:2 mixture of acetone and epon-araldite for at least 4 hours and up to 16 hours, throughout which time samples were kept on a rotator. On day two, samples were placed into freshly prepared epon-araldite plastic and set 58

PAGE 59

under a vacuum for 2 hours. Every 30 minutes during this 2 hour period, the vacuum was broken and resealed. Next, the plastic mixture was replaced for a third and final time, and samples were transferred into embedding molds and placed under a vacuum for an additional 60 minutes to remove air bubbles from the plastic. Embedding molds were then put in a 60 C oven for 12 hours. Embedded tissue was cut into ultra thin sections (80 nm) using a diamond knife. Sections were then collected on nickel grids and stained for TEM analysis with uranyl acetate (6 minutes) and lead citrate (5 minutes), and rinsed with distilled water. Morphometric and TEM Analysis Corpora lutea lengths were measured for each female (n=112). For morphometric analysis of luteal cells, one to four corpora lutea were selected from each female in the sample population, except female #5, which was not included in the morphometric analysis due to observed luteal cell regression that prohibited accurate measurement. Each CL was treated as a circle and divided into eight regions, with 20-30 luteal cells and surrounding endothelial cells and fibroblasts photographed per region at 1,000x. Measurements were taken from printed micrographs in order to facilitate cell analysis and data organization. A conversion factor was determined to change measurements collected from micrographs (mm) into micrometers (m), which represented the true cell size. This conversion was obtained by first measuring three different luteal cells with a calibrated digital micrometer (Leica Microsystem software for camera DFC 320), photographing (Leica DFC 320 microscope camera) and printing images of these same cells, and measuring their image from the micrograph (in mm). It was determined that dividing millimeter measurements by 2.4 resulted in a number that very closely approximated the true micrometer dimensions, so this conversion factor was applied to all micrographs generated at 1,000x. Cells from throughout the CL were examined so that any heterogeneity in cell distribution could be detected, as occurs in the manatees terrestrial 59

PAGE 60

relative the elephant [105]. Cell nuclei were measured if present at their full length, which was indicated by the presence of a dark ring along the perimeter of the nucleus. Lengths and widths of endothelial cell and fibroblast nuclei were also measured. A Hitachi H-7000 transmission electron microscope was used to analyze TEM specimens at the University of Floridas Electron Microscopy and Bioimaging laboratory, at magnifications ranging from 2,000x to 50,000x. Statistical Analysis Statistics were completed using SigmaStat (Systat Software, Inc., Richmond, California). An ANOVA on Ranks (Dunns method) was used to test differences in corpora lutea length among females (#1 n=22, #2 n=9, #3 n=8, #4 n=7, #5 n=20, #6 n=46) and a t-test was used to test differences in CL size between the left and right ovaries of each female (#1 n=16, 6; #2 n=4, 5; #4 n=2, 5; #5 n=8, 12; #6 n=21, 25). An ANOVA on ranks was used to test for significant differences in cell length and nucleus length between GLCs (cell n=1,015; nucleus n=1,157) and TLCs (cell n=58; nucleus n=70). To test differences in GLC length among females, an ANOVA on ranks was used (#1 n=89, #3 n=44, #4 n=480, and #6 n=402). An ANOVA on ranks was also used to test for significant differences in GLC length among the corpora lutea of individual females, in cases where more than one CL was examined. For female #1, two corpora lutea were examined, from which 165 and 69 GLCs were measured. Two corpora lutea from female #3 were also examined, with 50 and 44 GLCs measured, respectively. Three corpora lutea from female #4 were examined, from which 223, 227, and 130 GLCs were included in the statistical analysis. For female #6, 69, 141, 35, and 157 GLCs were measured from each of the four corpora lutea analyzed in the study. Changes in mean GLC length through diestrus and pregnancy were compared to changes in CL length (n=12) using a Pearson correlation. An ANOVA or ANOVA on ranks was also used to test significant difference in endothelial cell nuclear length among the corpora lutea of individual females (longitudinal sections: #1 n=2, #4 60

PAGE 61

n=3, #6 n=4) as well as among the studys females (longitudinal section: #1 n=53, #2 n=5, #3 n=3, #4 n=149, #6 n=352; cross section: #1 n=4, #4 n=22, #6 n=16). Both intra-female (number of corpora lutea examined: #1 n=2, #3 n=2, #4 n=3, #6 n=4) and inter-female (#1 n=24, #2 n=17, #3 n=12, #4 n=359, #6 n=156) differences in fibroblast nucleus length (longitudinal section) were also tested using ANOVA or ANOVA on ranks. For all statistical tests, P was determined to be significant at 0.05. Results The Manatee Corpus Luteum Active corpora lutea may be present on one or both ovaries, and five of six females in this study exhibited corpora lutea on both ovaries (#1, #2, #4, #5, and #6), while female #3 had corpora only on the right ovary. Of the four pregnant females in this study, three (#3, #4, and #5) presented a larger number of corpora lutea on the ovary ipsilateral to the uterine horn that contained the fetus, while female #6 exhibited a slightly higher number of corpora lutea on the ovary contralateral to the location of the fetus (31 vs. 25). The highest number of corpora lutea occurred in female #6 (n=56) and the lowest in female #4 (n=25) (Table 4-1). Each CL was surrounded by dense irregular connective tissue and connective tissue trabeculae containing fibroblasts. Capillaries lined with endothelial cells radiated from the periphery of each CL, towards its center. Newly formed corpora lutea were often observed with numerous capillaries traversing their serosal surfaces and were pale yellow on cut surfaces (Fig. 4-1). When corpora began to regress, cut surfaces appeared light gray to tan in color and microscopically, connective tissue became more prevalent, the numbers of luteal cells declined, and macrophages were often observed. Degeneration to a corpus albicans began within the central region of the CL and expanded outward, as evidenced in corpora lutea where lutein cells 61

PAGE 62

occurred peripherally and central regions contained only connective tissue (a connective tissue core). Within a single ovary or the two ovaries of an individual female, CL morphology sometimes varied. The left ovary of female #1 exhibited 11 corpora lutea with visible stigmata and one degenerating CL. Similarly, female #2 exhibited ruptured follicles in the left ovary and degenerating corpora lutea on the right ovary. Two of 23 corpora lutea on the right ovary of female #4 exhibited stigmata, although this female was in mid pregnancy (58 cm fetus) at the time of her death. Some corpora lutea of female #4 also exhibited connective tissue cores devoid of luteal cells. Finally, female #6, in late pregnancy, presented visible stigmata on one CL from the left ovary and two from the right ovary. Corpora lutea examined histologically ranged in length from 2.0-9.5 mm. The largest mean corpora lutea lengths were associated with late diestrus (8.2 mm) and the smallest with late pregnancy (4.5 mm). When corpora lutea from all six females were compared, corpora lutea from female #1 (early diestrus) and #3 (mid pregnancy) were significantly larger than those from two females in late pregnancy (#5 and #6) but no significant differences in CL size were found between the left and right ovary of each female (Table 4-1). Luteal Cells Granulosa lutein cells were most prominent and easy to distinguish within a CL. In manatee corpora lutea stained with H&E, these cells had a pink cytoplasm and granular, light purple nucleus. The TLC and its nucleus were approximately 50% of the size of a GLC and typically had darker cytoplasm and nuclei than GLCs. These differences in size and staining with H&E could be used to distinguish the two luteal cell types (Fig. 4-2). Granulosa lutein cells and TLCs occurred throughout the CL, but TLCs were more prevalent in peripheral regions. When GLC length and nucleus length were compared with TLC length and nucleus length for 62

PAGE 63

females #1, #4, and #6, TLCs were significantly smaller in cell and nucleus length (Table 4-2). These two lutein cell types were easiest to distinguish during mid pregnancy and most difficult to distinguish in late pregnancy due to observed changes in GLC size late in gestation. Granulosa lutein cells became reduced in size during late diestrus or pregnancy (Fig. 4-3). The greatest GLC length occurred in female #1 (CL B: 38.4 m) and the shortest mean was observed in a CL from female #6 (CL A: 14.7 m) (Table 4-3). Cells with lengths of approximately 20 m occurred in females #2 and #3, while female #4 had mean lengths of approximately 15 m. Mean GLC lengths for females #1 and #3 were significantly greater than those of females #4 and #6. The cell lengths of female #6 were also significantly greater than female #4. Granulosa lutein cell length was greater than cell width, but like length, width also decreased from a maximum in diestrus (#1B) to a minimum in late pregnancy (#6B). The GLC nuclear length decreased as the cell size was reduced, but the relationship was not as strong as for cell length. The GLC nuclei with the greatest mean length occurred in section #1B (9.5 m), while the smallest mean lengths were found in #4B and #6B (7.0 m), but nuclei with similar mean lengths occurred in females #1 and #6 (#1A: 8.4 m, #6D: 8.0 m). Nuclear width in GLCs ranged from 5.5 m (#6B) to 7.1 m (#1B and #2A). Where multiple corpora lutea were examined for a single female, the GLC lengths for each CL were compared. Significant differences were found among the corpora lutea of female #1 and female #6, where GLCs in section #6D were significantly longer than the mean lengths for #6A, B, and C. Additionally, the lengths of cells in CL #6C were significantly greater than #6A. No significant differences occurred for female #3 (P=0.611) or female #4 (P=0.223) (Table 4-3). Changes in GLC length through diestrus and pregnancy were compared to changes in CL length, with no significant correlation found (r=0.329, P=0.324). 63

PAGE 64

Binucleated GLCs (Fig. 4-4) were located throughout some corpora lutea of several females, including seven in female #1 (n=3 corpora lutea), six in female #2 (n=1), four in female #3 (n=2), 45 in female #4 (n=3), and three in female #6 (n=3). No TLCs could be distinguished in sections #1B, #2A, or #3A. The greatest mean TLC length occurred in female #3 (Table 4-4), but lengths were similar among all females, with the same mean length occurring in female #1(A) and female #4(A) (7.5 m). Theca lutein cell widths were similar for all females, typically ranging from 5.0-6.0 m, but the mean width for female #4(C) was 10 m. There were no differences in TLC length or width observed among females (P=0.08). Differences in TLC length and width among the CLs of individual females were not tested due to low sample size. Nuclear length ranged from 4.0 m (#3 B, #4B, and #6D) to 5.4 m (#4C) and width ranged from 2.9 m (#3B) to 5.0 m (#4C). Endothelial Cells and Fibroblasts Endothelial cell cytoplasm was very faint and difficult to visualize for measurement. The cell border could be seen on only two endothelial cells observed in longitudinal section and the mean length and width were 7.7 and 4.2 m, respectively. One endothelial cell could be measured in cross section, and had a cell length of 5.4 m and a width of 3.7 m. Both of these endothelial cells occurred in female #6 (section D). Endothelial cell nuclei were more well-defined than the cell membrane, were tubular in shape, and tapered slightly at each end (Fig. 4-5). These nuclei often appeared crescent-shaped due to their orientation within a capillary. For endothelial cell nuclei observed in longitudinal section, mean nuclear length ranged from 5.6 m (#6C) to 10.6 m (#1B) and nuclear width from 1.9 m (#4C) to 2.9 m (#1B). Where differences in endothelial cell nucleus length (longitudinal sections) could be tested among the corpora lutea of individual females, all females exhibited significant differences (#1, #4, and #6) (Table 4-5). Longitudinal sections of endothelial cell nuclei from 64

PAGE 65

female #6 were significantly smaller than those from females #1, #3, and #4 (Table 4-6). Endothelial cells observed in cross section had mean maximum nuclear diameters ranging from 3.0 m (#4C) to 5.4 m (#1B) and mean minimum nuclear diameters from 2.1 m (#4B) to 4.8 m (#1B). Fewer endothelial cell nuclei were distinguishable in cross section than in longitudinal section, and no endothelial cells were observed in section #3B. For endothelial cell nuclei in cross-section, significant differences were observed between female #1 and females #4 and #6, while there was no significant difference between #4 and #6. Finally, the greatest number of distinguishable endothelial cells observed (longitudinal and cross sections) occurred in female #6 (n= 368) while the lowest number occurred in female #2 (n=5) and female #3 (n=3) (Table 4-6). The cytoplasmic borders of fibroblasts were also very difficult to distinguish, but cell lengths were distinguishable in two corpora lutea examined. Section #4B contained five fibroblasts where cell length (9.4 m) and width (4.1 m) were measured and a mean determined. Also, one fibroblast in section #6D had a cell length of 5.4 m and a cell width of 2.9 m. Fibroblast nuclei were often thinner and more tapered than endothelial cell nuclei (Fig. 4-5). The greatest mean fibroblast nuclei length occurred in sections #1B (12.2 m) and #2A (12.7 m) and were shortest in #4B (6.9 m), while mean nuclear widths ranged from 2.3 m (#3B) to 3.8 m (#1B) (Table 4-7). One fibroblast observed in cross section from female #4B had a maximum nuclear diameter of 4.2 m and a minimum nuclear diameter of 3.3 m. The greatest numbers of fibroblasts were observed in corpora lutea from female #4, followed by female #6. Females #1 and #3 had similar numbers of fibroblasts, and the fewest fibroblasts occurred in female #2. There was only one significant difference when pairwise comparisons of fibroblast nucleus length were compared among females: the nuclei of female #1 were 65

PAGE 66

significantly longer than those of female #4 (Table 4-8). When corpora lutea within individual females were tested for difference in fibroblast nucleus length, no differences were found within females #3, #4, or #6, but a significant difference was observed between the two corpora lutea of female #1 (Table 4-7). Degenerative Changes Observed with TEM Upon examination with TEM, the corpora lutea from a mid pregnancy female (#4) were found to be undergoing degeneration. Luteal cells lacked organization and possessed few identifiable organelles (Fig. 4-6). Dense accumulations of free protein granules were present throughout the cytoplasm of GLCs and some luteal cells also contained low amounts of lipid droplets or multiple osmiophilic droplets. Clustered mitochondria were present and were the predominant organelle observed. These mitochondria sometimes retained small numbers of cristae, but many others lacked discernible cristae and had a vesicular appearance. Brief expanses of rough endoplasmic reticulum (RER) were visible in some GLCs. Nuclei contained heterochromatin of varying amounts, sometimes scattered throughout the nucleus and other times accumulated at its edge (Fig. 4-7). Some GLC nuclei appeared almost white in color because of the absence of nuclear material, while others were almost completely black due to the dense accumulation of heterochromatin. Cells with pleomorphic nuclei lined with a dense layer of heterochromatin were observed, which were likely macrophages. Infiltration of collagen in the intercellular spaces of the CL was also observed. Endothelial cell and fibroblast nuclei also exhibited dense accumulation of heterochromatin around the periphery. Some normal capillaries were observed, but others exhibited partial occlusion of the lumen by distended endothelial cells. 66

PAGE 67

Discussion The Manatee Corpus Luteum While samples from only a limited number of females were available for analyses, the presence of multiple corpora lutea within the ovaries of individual females helped to greatly increase sample size. It was hypothesized that the number and size of manatee corpora lutea would vary at different reproductive stages, with more corpora present during pregnancy than diestrus and a possible decline in the number of corpora by late pregnancy. While a variation in the number of corpora lutea was observed, ranging from 25 and 26 in mid pregnancy to 49 and 56 in late pregnancy, the trend differed from the expected pattern, with the number of corpora lutea increasing in late pregnancy. However, these late pregnancy corpora lutea were significantly smaller than those observed during diestrus, which implies a possible decline in the function of individual CL prior to parturition. The observance of maximum mean CL lengths during diestrus may represent a necessary accumulation of luteal tissue for progesterone production in the event of pregnancy. The observed numbers of corpora lutea were independent of manatee total length, suggesting that the differences were not related to animal age, as has been suggested for elephants [106,107]. It is possible that the observed increase in CL number during late pregnancy is a mechanism for maintaining progesterone secretion as the function of each individual CL decreases. For those females with corpora lutea on both ovaries, no significant difference in CL length was found between the left and right ovary, regardless of fetal position, which suggests that structures on both ovaries contribute to progesterone production in support of pregnancy. It was also found that like African elephants [108], the number of corpora lutea on each ovary cannot be used to determine the location of the fetus, as female #6 exhibited a higher number of corpora lutea on the ovary contralateral to the pregnant uterine horn (n=31 vs. 25). 67

PAGE 68

Manatee corpora lutea (2.0 to 9.5 mm in histological sections) are smaller than those of similarly sized domestic species such as equines and bovines, where the primary CL measures approximately 20 mm in length [109,110]. While manatee corpora lutea are slightly larger prior to fixation and histological processing, they are still not as large as those of equines and bovines. However, the corpora lutea of the dugong (1 to 14 mm) [111] are similar in size to those of the manatee. In African elephants, one CL is larger than the others, and contains nearly 50% of the luteal tissue [112], but all manatee CLs within individual females were approximately equal in size. Stigmata were observed on several CLs of females in mid (#4) and late (#6) pregnancy. Historical reports of the horse suggested that the mare might ovulate during gestation to form accessory corpora lutea (versus luteinization of unovulated follicles) [113-115], but such a strategy seems unlikely for the horse, manatee, or other species. It is possible that although these structures in the manatee appeared to be recently formed accessory corpora lutea, they were simply corpora lutea whose development was arrested during diestrus. However, there is other evidence to suggest that manatee corpora lutea form asynchronously. Such findings include luteal cell proliferation within a mid pregnancy female (#3), which has also been observed in the tree hyrax [116]. Additionally, the presence of large vesicular (pre-ovulatory) follicles in the ovaries of female #6 (late pregnancy), and observance of degenerating corpora lutea in females #1, #2, #4, and #6 indicates that both development and degeneration of manatee corpora lutea are asynchronous. Additional information about the manatees in this study, gathered prior to or during necropsy, provided insight into the manatee estrous cycle and breeding. First, attempts by male manatees to mate with female #1 just prior to her rescue suggest that the observed luteal 68

PAGE 69

development in this female corresponded to a receptive mating period. Additionally, female #2, in late diestrus, was lactating and exhibited a placental scar in the left uterine horn. These observations indicate that manatees can return to estrus prior to complete weaning of calves, which supports an earlier hypothesis by other authors, based on observations of mating herds of males following females with larger calves [1]. Luteal Cells It was hypothesized that two cell types could be distinguished microscopically in the manatee, and this was confirmed. Both GLCs and TLCs were observed, with GLCs being much more abundant than TLCs. While this pattern is also common in some other species, TLCs do contribute significantly to the composition of the CL in bovines, porcines, and ovines [91]. In bovines, large luteal cells (equivalent to GLCs) contribute 3.0% of the steroidogenic cells, while small luteal cells (TLCs) contribute 26% [117,23]. In ewes, small luteal cells outnumber large luteal cells with a ratio of approximately 5:1 [118,119]. The current results suggest that TLCs may contribute no more than 4% of manatee luteal cells, based on the numbers of these cells observed in diestrus and pregnancy. In other species, GLCs contribute significantly to progesterone production [120-122], and in bovines, these cells produce 20-times as much progesterone as equal numbers of TLCs [89]. If true for the manatee, this could explain why GLCs are present in such high numbers, particularly if individual manatee corpora lutea have low steroidogenic potential, as has been previously suggested [104]. Granulosa lutein cell lengths of 16-25 m, 23-35 m, and 24-45 m have been reported for hyraxes, ovines, and bovines, respectively, with TLC lengths of 11.4 m, 12-22 m, and 23 m [23,87,116,123]. The mean manatee GLC (19.3 m) and TLC (8.5 m) lengths were slightly smaller than the species mentioned above. However, many manatee corpora lutea examined in this study were from later stages of pregnancy, and were found to be significantly smaller than 69

PAGE 70

those from diestrus. Thus, the lengths of GLCs in healthy, fully functional corpora lutea are likely greater than many lengths observed in this study. A previous manatee study did not describe any changes in lutein cell size throughout pregnancy [104], although this has been reported in African elephants [108,120], and is believed to occur in the manatee. Additionally, although the observed decreases in CL and GLC length were not significantly correlated, it is believed that this is due to the small number of corpora lutea tested (n=12), and that a larger sample size would have shown that the two are in fact correlated. It was possible to distinguish GLCs and TLCs based on size (TLCs significantly smaller than GLCs), even though a previous study indicated that primary luteal cells and secondary luteal cells could not be distinguished, as both ranged in length from 13-27 m [104]. In addition to differences in size, the TLCs stained more darkly than the GLCs with H&E and were more elongate in shape than the more spherical to polyhedral GLCs. The present finding that manatee TLCs are more common peripherally contrasts an earlier study which stated that the two cell types are distributed fairly homogenously throughout the CL [104], as occurs in most mammals [21]. While there was no clear spatial separation of the two cell types observed in this study, or clustering of different types of luteal cells, as reported in the African elephant [105], TLCs were observed more frequently in the peripheral regions of manatee corpora lutea. This may have been due to the fact that TLCs do not contribute significantly to the cell population of the manatee corpus luteum, so they do not proliferate as extensively throughout the corpus luteum or within a particular region of the corpus luteum as they do in species where they are more prevalent. Binucleated GLCs were observed, and have also been reported in the African elephant [108]. The greatest number were observed in female #4 (mid pregnancy) and were nearly absent 70

PAGE 71

in female #6 (late pregnancy). Female #4 had been struck by a watercraft 24-48 hours prior to her death, and based upon the regression of corpora lutea observed in this study, was believed to have been undergoing abortion at the time of her death. It has been found that enhanced numbers of binucleated cells in liver hepatocytes are related to the severity of hepatocellular carcinoma [124]. While the observance of binucleated GLCs in female #4 is likely unrelated to her abortion, it may be possible that binucleated cells represent a corpus luteum of low quality. Because manatees form multiple corpora lutea during diestrus, cell division may sometimes be inefficient, resulting in such flaws as cells that go undivided and become binucleated luteal cells. Perhaps it is even possible that binucleated cells could have physiological consequences within the corpora lutea, related to such important processes as progesterone production. Both luteal cell types exhibited centrally located nuclei and a nucleolus was often visible. Both GLC and TLC nuclei were elliptical in shape, as their length was consistently greater than their width. Manatee GLC nuclei were previously described as spherical, while TLC nuclei were identified as spindle-shaped [104]. While GLCs in this study often did appear more rounded than nuclei of TLCs, no nuclei were observed which could be described as spherical. The mean nuclear length for GLCs in this study was 7.8 m, with a range of 7.0-9.5 m. These values are similar to the range of 6.0-9.0 m previously reported in manatees [104], as well as values in hyraxes: 8.0 m and 7.7-10.0 m [116,123]. By comparison, the theca lutein cell mean nucleus length was 4.5 m, with a range from 4.0-5.4 m, which was much smaller than the range of 6-16 m reported in the previous manatee study [104]. The present finding that manatee TLCs are more common peripherally contrasts an earlier study which stated that the two cell types are distributed fairly homogenously throughout the CL [104], as occurs in most mammals [21]. While there was no clear spatial separation of the two 71

PAGE 72

cell types observed in this study, or clustering of different types of luteal cells, as reported in the African elephant [107], TLCs were observed more frequently in the peripheral regions of manatee corpora lutea. This may have been due to the fact that TLCs do not contribute significantly to the cell population of the manatee corpus luteum, so they do not proliferate as extensively throughout the corpus luteum or within a particular region of the corpus luteum as they do in species where they are more prevalent. With regard to the longevity of manatee corpora lutea, it was hypothesized that luteal cells would decrease in size late in diestrus or pregnancy, indicating their regression. The degree to which corpora lutea regress prior to parturition varies among species, and in some mammals, including the mare, is facilitated by secretion of necessary hormones from the placenta [79]. A previous manatee study did not describe any changes in lutein cell size throughout pregnancy [104], although this has been reported in African elephants [108,120]. The significant decline in luteal cell size by late pregnancy (female #6) observed in current study suggests that regression, at least for some corpora lutea, begins during gestation. It is likely that not all corpora lutea regress simultaneously, since significant differences in GLC length were observed among the corpora lutea of female #6. This suggests that some corpora lutea were either not regressing, or regressing at a different rate. Although the observed decreases in CL and GLC length were not significantly correlated, it is believed that this is due to the small number of corpora lutea tested (n=12), and that a larger sample size would have shown that the two are in fact correlated. The decrease in luteal cell size in late pregnancy may also be associated with the observed increase in corpus luteum number in late pregnancy, where these additional corpora lutea are required to aid in progesterone production in manatees where earlier formed corpora lutea have begun to regress. 72

PAGE 73

Endothelial Cells and Fibroblasts Endothelial cells and fibroblasts were not morphologically described, quantified, or related to CL age in a previous study of the manatee CL [104]. In the present study, it was hypothesized that trends in the number of observed endothelial cell and fibroblast nuclei would be related to manatee reproductive status, where endothelial cells (and associated capillaries) would be characteristic of healthy, fully functional corpora lutea, and fibroblasts would be associated with regressing corpora lutea. In the manatee, the greatest numbers of endothelial cell nuclei were observed during late pregnancy, while the fewest were observed during late diestrus and mid pregnancy. This finding was contrary to our hypothesis, and was surprising, considering that healthy luteal cell function is characterized by a well-connected network of vasculature [125,126] to transport progesterone that is secreted from luteal cells. For this reason, it was expected that the greatest number of endothelial cells would be associated with diestrus and a significant portion of pregnancy, with a loss of vascularization accompanying corpora lutea regression [90] in late gestation, or post-partum. In this study, the highest numbers of endothelial cells were observed in the corpora lutea of female #6, in late pregnancy. This could indicate ample progesterone production and transport within the corpora lutea, despite the diminished luteal cell size. However, it is also possible that vascular regression [103] is not one of the primary mechanisms of corpus luteum regression in the manatee. Finally, this trend may be explained by the fact that endothelial cells were more difficult to distinguish and measure from diestrus through mid pregnancy, due to the close proximity of robust luteal cells, which dominated the CL during these stages. As hypothesized, fibroblast nuclei were most prevalent in corpora lutea that were regressing (female #4 [abortion] and female #6 [late pregnancy]) and occurred much less frequently in younger corpora lutea. However, the longest fibroblast nuclei were associated with 73

PAGE 74

diestrus, even though the number of fibroblast nuclei was low at this stage. This finding indicates that fibroblast nucleus size also decreases during luteal regression, even as the prevalence of these cells increases. Degenerative Changes Observed with TEM It was hypothesized that transmission electron microscopy could be used to confirm luteal degeneration in a female suspected to be undergoing abortion (#4). This hypothesis was proven using comparative indicators of regression in luteal cells and organelles from other species, including the absence of accumulated osmiophilic lipid droplets to indicate the end of steroid hormone production but the continued accumulation of lipid [98,100,101]. The low numbers of lipid droplets in the CL of female #4suggest that the traumatic injury caused an abrupt end to all steroid producing processes in the ovaries. Osmiophilic granules were observed in the corpora lutea of female #4, and while similar granules in other species have been found to contain progesterone [97] and relaxin [101,127], it is unlikely that the granules in the manatee contained steroid hormone. This is because a loss of such granules is an early indicator of regression in other species [97,100,102], and these manatee corpora lutea were in an advanced stage of regression. However, it is possible that these granules may have contained an enzyme involved in the luteolytic process. The amount of collagen observed in the inter-luteal spaces also suggest advanced regression [23,95]. The swollen appearance of some endothelial cells, which caused them to protrude into capillaries, is also associated with degeneration in other species [95,97,102,103], as it reduces blood flow through the corpus luteum. Nuclei of luteal cells as well as endothelial cells and fibroblasts often contained dense accumulations of heterochromatin. While diffuse heterochromatin characterizes healthy GLCs, and healthy TLC nuclei may be bordered by heterochromatin, the observed accumulation far exceeded normal conditions 74

PAGE 75

described for other species [23,120]. This heterochromatin accumulation is believed to characterize degeneration in the manatee, as it does for the mare [95]. Conclusion This study provided information about the manatee CL, including a general description of its structure, identification of its luteal cell types, documentation of fibroblast and endothelial cell morphology, and the first transmission electron micrographs of manatee luteal cells. For the first time, intra-female and inter-female differences in corpus luteum and luteal cell morphology have been documented, and related to changes in corpus luteum function. These findings suggest that regression of at least some of the manatees numerous corpora lutea begins during pregnancy. However, development of accessory corpora lutea during gestation may assist with progestin development in support of pregnancy. Degenerative changes in manatee corpora lutea were also identified and included a loss of cellular organization, few identifiable organelles, clustered mitochondria with few cristae, abundant nuclear heterochromatin, and stromal collagen. These results provide new information about the manatees poorly understood reproductive trends and represent progress towards understanding the role of multiple corpora lutea in this species. 75

PAGE 76

Table 4-1. Corpora Lutea (CL) Morphometrics for Study Specimens ID Status Total Length (cm) Left Ovary Mean CL Length SD; Range (mm) (n measured / total) Right Ovary Mean CL Length SD; Range (mm) (n measured/ total) Total CL # CL: Left v. Right Ovary (t-test) #1 a Early diestrus 345 6.6 1.5; 3.0-9.0 (n=16/17) 7.2 1.7; 5.5-9.5 (n=6/8) 34 NSD P=0.438 #2 Late diestrus 263 8.2 0.5; 8.0-9.0 (n=4/4) 6.1 1.9; 4.5-9.0 (n=5/32) 36 NSD P=0.071 #3 a Mid pregnancy, 39 cm fetus 275 ---(n=0/0) *7.4 1.4; 5.0-9.0 (n=8/26) 26 No Test #4 Mid pregnancy, 58 cm fetus 300 6.3 1.1; 6.0-8.0 (n=2/2) *5.2 0.3; 5.0-8.0 (n=5/23) 25 NSD P=0.262 #5 b Late pregnancy, 112 cm fetus 281 *4.5 1.2; 2.0-6.0 (n=8/38) 5.0 1.0; 3.5-6.5 (n=12/18) 49 NSD P=0.252 #6 b Late pregnancy, 131 cm fetus 334 *5.2 0.9; 4.0-7.0 (n=21/25) 4.9 1.0; 3.0-6.5 (n=25/31) 56 NSD P=0.303 a,b Denote significant differences in CL length. The pregnant uterine horn is denoted by*. Table 4-2. Determination of Significant Differences in Mean GLC and TLC Lengths Mean GLC Length m (n) Mean TLC Length m (n) P-value #1 Cell Length 45.3 (89) 18.1 (8) <0.001 #1 Nucleus Length 20.1 (68) 11.1 (8) <0.001 #3 Cell Length 53.2 (44) 28.9 (2) --#3 Nucleus Length 18.9 (47) 9.5 (2) --#4 Cell Length 36.6 (480) 19.6 (31) <0.001 #4 Nucleus Length 17.0 (609) 10.4 (42) <0.001 #6 Cell Length 39.2 (402) 21.3 (17) <0.001 #6 Nucleus Length 17.9 (433) 11.2 (18) <0.001 No TLCs were observed in female #2. No statistics are presented for female #3 due to the low number of TLCs observed. 76

PAGE 77

Table 4-3. Mean Measurements and Ranges of GLCs ID# Cell Length, Range m (n) Cell Width, Range m (n) Nucleus Length, Range m (n) Nucleus Width, Range m (n) 1A 19.8, 10.8-33.3 (165) a 13.7, 7.1-28.3 (164) 8.4, 5.0-11.2 (128) 6.8, 3.7-9.6 (129) 1B 38.4, 14.2-60.8 (69) b 14.3, 6.7-28.3 (80) 9.5, 6.2-11.7 (91) 7.1, 5.0-10.4 (93) 2A 19.6, 10.0-41.2 (142) 12.2, 5.8-22.5 (147) 8.9, 4.6-12.9 (145) 7.1, 4.2-11.2 (146) 3A 21.5, 9.6-43.3 (50) aa 11.0, 5.8-20 (53) 8.0, 4.6-11.7 (50) 6.2, 3.7-11.7 (51) 3B 22.2, 12.1-37.1 (44) aa 11.6, 7.1-21.7 (46) 7.9, 5.4-11.2 (47) 6.5, 4.2-8.7 (47) 4A 15.2, 8.3-29.6 (281) aaa 10.3, 5.4-20.8 (289) 7.1, 4.2-11.2 (357) 6.1, 3.3-9.6 (357) 4B 15.3, 7.5-27.9 (308) aaa 10.0, 4.6-20.8 (322) 7.0, 4.6-12.9 (357) 5.9, 2.9-9.6 (358) 4C 15.4, 8.3-37.5 (131) aaa 9.9, 4.6-21.2 (134) 7.0, 4.2-12.5 (148) 5.9, 2.9-9.2 (148) 6A 14.7, 7.5-37.5 (135) aaaa 9.8, 4.6-21.7 (140) 7.2, 4.2-10.0 (146) 6.0, 3.3-8.7 (147) 6B 15.5, 8.3-32.9 (157) aaaa 9.1, 5.0-20.0 (161) 7.0, 5.0-10.0 (166) 5.5, 3.7-8.3 (166) 6C 16.4, 9.2-32.9 (141) aaaa 10.5, 5.4-17.5 (145) 7.6, 5.4-12.1 (152) 6.2, 2.5-10.0 (151) 6D 18.2, 9.2-36.2 (143) bbbb 11.1, 6.2-20.0 (157) 8.0, 5.8-11.7 (169) 6.5, 4.2-10.0 (169) a,b The GLCs in CL 1A a were significantly shorter than those in 1B b aa There was no significant difference in GLC length between the two CLs of female #3. aaa No significant differences in GLC length were observed among the three corpora lutea examined for female #4. aaaa, bbbb The GLCs in corpus luteum 6D bbbb were significantly larger than those from 6 A, B, and C. Table 4-4. Mean TLC Measurements Cell Length m (n) Cell Width m (n) Nucleus Length m (n) Nucleus Width m (n) #1A 7.5 (n=8) 5.9 (n=8) 4.6 (n=8) 3.7 (n=8) #3B 11.7 (n=2) 6.0 (n=2) 4.0 (n=2) 2.9 (n=2) #4A 7.5 (n=14) 5.6 (n=15) 4.7 (n=19) 3.7 (n=19) #4B 8.6 (n=16) 5.4 (n=16) 4.0 (n=22) 3.2 (n=22) #4C 10.8 (n=1) 10.0 (n=1) 5.4 (n=1) 5.0 (n=1) #6A 6.9 (n=2) 5.6 (n=2) 4.2 (n=2) 4.0 (n=2) #6B 10.0 (n=8) 5.9 (n=8) 5.2 (n=8) 4.0 (n=8) #6C 6.8 (n=5) 6.3 (n=5) 4.2 (n=5) 3.5 (n=5) #6D 7.1 (n=2) 6.2 (n=2) 4.0 (n=2) 3.3 (n=2) No significant differences in TLC length or width were observed among the studys females. 77

PAGE 78

Table 4-5. Mean Endothelial Cell Measurements Section Longitudinal Section Nucleus Length m (n) Longitudinal Section Nucleus Width m (n) Cross Sectional Mean Maximum Nuclear Diameter m (n) Cross Sectional Mean Minimum Nuclear Diameter m (n) #1A 6.1 (n=38) a 2.4 (n=38) 4.2 (n=2) 3.3 (n=2) #1B 10.6 (n=15) b 2.9 (n=15) 5.4 (n=2) 4.8 (n=2) #2A 7.0 (n=5) 2.7 (n=5) ----#3A 7.9 (n=3) 2.2 (n=3) ----#4A 7.1 (n=38) aa 2.6 (n=38) ----#4B 6.7 (n=69) 2.5 (n=70) 3.1 (n=14) 2.1 (n=14) #4C 5.8 (n=42) bb 1.9 (n=42) 3.0 (n=7) 2.2 (n=7) #6A 5.7 (n=85) aaa 2.2 (n=85) ----#6B 7.4 (n=29) bbb 2.5 (n=29) 3.5 (n=2) 2.9 (n=2) #6C 5.6 (n=151) aaa 2.4 (n=153) 3.2 (n=10) 2.7 (n=10) #6D 5.8 (n=85) aaa 2.4 (n=85) 3.7 (n=4) 2.9 (n=4) Longitudinal sections of endothelial cell nuclei were significantly smaller in #1A a than #1B b Longitudinal sections of endothelial cell nuclei were significantly longer in #4A aa than #4C bb Longitudinal sections of endothelial cell nuclei were significantly smaller in #6 A, C, and D aaa than in #6B bbb Table 4-6. Endothelial Cell Nuclei Summary: Longitudinal (LS) and Cross Sections (XS) Female # Sample Size (n) Mean SD (m) Range (m) #1 LS 53 7.4 2.7 a 2.9-13.3 #1 XS 4 4.8 1.0 1 4.2-6.2 #2 LS 5 7.0 1.7 0.8-9.6 #3 LS 3 7.9 0.8 a 7.1-8.7 #4 LS 149 6.5 2.0 a 2.5-12.5 #4 XS 22 3.3 0.7 2 1.7-4.6 #6 LS 350 5.8 1.9 b 2.1-14.6 #6 XS 16 3.4 0.8 2 2.1-5.0 Longitudinal sections of endothelial cells were significantly longer in females #1 a #3 a and #4 a than in #6 b Cross-sections length of endothelial cell nuclei was significantly greater in female #1 1 than in females #4 2 and #6 2 78

PAGE 79

Table 4-7. Mean Fibroblast Nuclei Longitudinal Sections Ovarian Section Number Nucleus Length m (n) Nucleus Width m (n) 1A 7.4 (n=10) a 3.7 (n=10) 1B 12.2 (n=14) b 3.8 (n=14) 2A 12.7 (n=5) 3.7 (n=5) 3A 8.3 (n=13) 2.5 (n=13) 3B 9.0 (n=11) 2.3 (n=11) 4A 8.8 (n=218) 2.9 (n=217) 4B 6.9 (n=97) 2.5 (n=97) 4C 8.9 (n=44) 2.5 (n=44) 6A 9.1 (n=40) 2.7 (n=40) 6B 9.2 (n=38) 2.9 (n=38) 6C 9.0 (n=25) 2.6 (n=25) 6D 8.6 (n=53) 2.7 (n=53) Fibroblast nuclei were significantly smaller in corpus luteum A a than in corpus luteum B b of female #1. Table 4-8. Fibroblast Nucleus (Longitudinal Sections) Summary Data Female Number N Mean SD (m) Range (m) 1 24 10.2 3.1 a 5.4-15.8 2 5 9.6 2.6 6.2-14.6 3 23 9.0 1.7 6.2-12.1 4 359 8.7 1.8 b 5.0-15.8 6 156 9.0 1.9 5.4-15.0 Fibroblast nuclei were significantly longer in female #1 a than female #4 b 79

PAGE 80

A B H C M 2 1 Figure 4-1. Gross sections of manatee corpora lutea. A) Two well-vascularized corpora lutea (1,2) from a diestrus female Florida manatee (#1), surrounded by vesicular follicles. Both corpora are <1 cm in length. B) Cross-section of a 9 mm corpus luteum, with surrounding cortex and medulla attached. The corpus is still developing as indicated by remnants of hemorrhage within the structure (arrow). A and B scale bars are in mm. C: cortex; M: medulla; H: hemorrhage. 80

PAGE 81

GLC TLC Figure 4-2. Appearance of GLCs versus a TLC. Two granulosa lutein cells can be seen on the left side of this micrograph, and a single theca lutein cell is visible on the right. Theca lutein cells are approximately 50% of the size of granulosa lutein cells and have a more darkly stained cytoplasm and nucleus. GLC: granulosa lutein cell; TLC: theca lutein cell (1,000x) 81

PAGE 82

A B C D E F Figure 4-3. Micrographs A-F illustrate the microscopic appearance of granulosa lutein and support cells of corpora lutea from each of the studys six females. The largest GLCs occur in female #1 (A), while the smallest occur in #4 (D) and #6 (F). There is a clear difference in cell size between early (A) and late (B) diestrus. The cells of #2 (B) and #3 (C) are intermediate in size. The presence of connective tissue and fibroblasts in the corpus luteum is evident in females #3-#6 (C-F), and particularly evident in female #4 (D). All cells other than those from female #1 (A) appear shrunken in size and lack contact between adjacent cells. A binucleated cell is visible in female #6 (F) (arrow). Nuclear size is relatively consistent among all six females. (1,000x) 82

PAGE 83

Figure 4-4. A binucleated granulosa lutein cell from female #1, designated with a star, is surrounded by uninucleated lutein cells. (400x) Fibroblasts GLC Capillary with erythrocytes Endothelial Cells Figure 4-5. Endothelial cells and fibroblasts are visible in a section of a corpus luteum from female #2 (late diestrus). Granulosa lutein cells and capillaries are also visible in the section. (400x) 83

PAGE 84

L N HC P Figure 4-6. Cross section of a degenerating granulosa lutein cell from female #4 observed with TEM. This luteal cell is characterized by disorganization and degeneration of organelles. The nucleus (N) contains sparse heterochromatin (HC) and protein granules (P) and a lipid droplet (L) can be distinguished in the cytoplasm. (3,500x) N M P RER Figure 4-7. Higher magnification view of organelles from Figure 4-6. The edge of the nucleus (N) is visible in the upper right hand corner and rough endoplasmic reticulum (RER), a mitochondrion (M), and protein granules (P) are present in the cytoplasm. Vesicular structures surrounding the labeled mitochondrion are likely mitochondria that have lost their cristae. (15,000x) 84

PAGE 85

CHAPTER 5 ASSESSMENT OF THE STEROIDOGENIC POTENTIAL OF FLORIDA MANATEE FOLLICLES AND CORPORA LUTEA USING STEROIDOGENIC ACUTE REGULATORY PROTEIN Introduction The ovaries of manatees in estrus are characterized by numerous, small vesicular follicles (<1 cm diameter), while variable numbers of corpora lutea (14 to 58) occur on one or both ovaries during diestrus and pregnancy [128]. Other monotocous species, such as bovines [129], typically produce one dominant follicle during estrus, which becomes a corpus luteum. In canines, which are polytocous, the number of corpora lutea corresponds to the number of offspring associated with each pregnancy [130]. Elephants, however, which are close relatives of manatees [7,8], also produce more than one follicle and corpus luteum during each estrous cycle, but give birth to a single offspring [62,105]. The advantage to manatees of producing a high number of follicles and corpora lutea is unknown. A previous study hypothesized that individual corpora lutea may have low steroidogenic potential, necessitating a higher number of structures to maintain pregnancy [104], but this hypothesis was not previously tested. Luteal progestin concentrations in the manatee have not been studied, but serum progesterone concentrations have been measured, and are lower than those observed in many domestic species (see Chapter 2), particularly during pregnancy [131-133]. This apparent disconnect between the number of corpora lutea and circulating progesterone concentrations in the manatee has not previously been investigated, but knowledge of the steroidogenic potential of manatee corpora lutea could help explain the patterns of follicular and luteal development observed in this species. Steroidogenic acute regulatory protein (StAR) controls the movement of cholesterol across the outer mitochondrial membrane and into the inner membrane so it can be converted into 85

PAGE 86

steroid hormones [134]. In this manner, StARs ability to bind to and transport cholesterol controls the rate of steroid hormone formation [135]. Because of this relationship, immunohistochemical detection of StAR protein in luteal cells can be used as an indicator of their ability to produce steroids. Since progesterone is the dominant secretory product of all luteal cells, StAR expression can be used as a measure of progesterone production within corpora lutea, with reduced StAR expression indicating less progesterone production [136]. The objectives of this study were to: (1) identify StAR localization within manatee follicles and corpora lutea, including regional differences; and (2) investigate intra-animal and (3) inter-animal variation in StAR localization to help determine the steroidogenic potential of manatee follicles and corpora lutea. Materials and Methods Sample Collection Florida manatee ovarian tissue was obtained from the Florida Fish and Wildlife Conservation Commissions Marine Mammal Pathobiology Laboratory in St. Petersburg, Florida and analyzed at the University of Florida. The ovaries were removed at the time of necropsy, and pregnancy was assessed by locating a fetus in one of the uterine horns. Where a fetus could not be detected, corpora lutea observed on the ovaries were said to be associated with a non-pregnant diestrus. The presence of large vesicular follicles in the ovaries was used to define estrus. Tissues were collected in 10% NBF and transferred to PBS after 24 hours. Sample Population Tissues from four adult females (>265 cm straight length [50]) were utilized for StAR immunohistochemistry. Three of the carcasses were assessed as fresh at the time of necropsy, and one was moderately decomposed. Stage of pregnancy for adult females was based on fetal length, where early (<31 cm total straight length), mid (31-80 cm), and late (>80 cm) gestation 86

PAGE 87

were defined. These thresholds were based upon observations that calves may be born ranging in length from 82 to 160 cm or 95 to 155 cm [1,24] and an assumption that fetal growth is approximately linear. The number of specimens tested immunohistochemically for each female varied based upon the number of structures observed at necropsy. Female #1, with no grossly visible vesicular follicles and a small number of developing follicles observed microscopically, was assessed as anestrous, and one ovarian cross-section was examined. Female #2 was in early diestrus and 19 different corpora lutea were examined. Four corpora lutea were tested for female #3 in mid pregnancy (58 cm) and 15 different corpora lutea were examined for female #4 in late pregnancy (131 cm). Knowledge that female #3 died 24-48 hours after being struck by a watercraft, as well as morphological assessment of the corpora lutea using light and transmission electron microscopy (see Chapter 4), indicated that female #3 was likely undergoing abortion at the time of her death. Immunohistochemistry For immunostaining, sections embedded in paraffin were cut at 5 m and mounted on SuperFrost/Plus positively charged microscope slides (Fisher Scientific). Additional sections were also cut and mounted for histological examination with H&E. All slides were placed in a 60 C oven for 1-2 hours. Slides were deparaffinized by placement in three changes of xylene (5 minutes each), three changes of 100% EtOH, two changes of 95% EtOH, two changes in tap water, and two changes in distilled water (all 2 minutes each). Slides were then placed in PBS for 2 minutes, followed by quenching with 0.3% H 2 O 2 for 20 minutes to block endogenous peroxidase activity. After removing excess PBS, slides were treated with 5% normal horse serum for 30 minutes to block non-specific binding of the polyclonal primary antibody. Slides were then rinsed again with PBS. Next, the primary antibody (rabbit polyclonal to rat StAR [Abcam Inc., Cambridge, MA]) 87

PAGE 88

was applied to slides at a dilution of 1:200, after having tested concentrations from 1:100 to 1:1000. Slides were incubated overnight within a humidified changer in a walk-in cooler. On day 2, a biotinylated second antibody to rabbit (Invitrogen, Carlsbad, CA), was applied. Sections were washed in two, 2 minute changes of PBS, and then the biotinylated link was added for 10 minutes. Following additional washes in PBS, streptavidin-HRP (horseradish peroxidase) was applied to slides for 10 minutes and the AEC chromagen was applied to the slides for 5 minutes. Finally, slides were washed in two, 2 minute changes of distilled water, and cover slipped with glycerol jelly. A negative control slide was created by eliminating primary antibody, and a positive control was determined by observed tissue staining in desired regions. Staining Analysis Differences in StAR staining within individual follicles and corpora lutea were evaluated and sections were photographed at 20x, 100x, 250x, and 400x to document stain intensity. Results were examined after the methodology used in Pollack et al. [137], where staining was evaluated semi-quantitatively (absent: 0; minimal: 1; moderate: 2; intense: 3). Intra-animal and inter-animal differences in staining were evaluated numerically by tallying the staining scores and dividing by the number of sections examined, to calculate a mean staining score. Statistical Analysis Statistics were completed using SigmaStat (Systat Software, Inc., Richmond, California). Significant differences in follicle staining between the left and right ovaries were tested for females #3 (n=19) and #4 (n=65) using a one-way ANOVA on ranks. This test was also used to check for differences in corpus luteum staining between ovaries for females #2 (n=19), #3 (n=4), and #4 (n=15). Inter-female differences in follicle (n=131) and corpus luteum (n=38) expression of StAR were tested with a one-way ANOVA on ranks (Dunns method). Statistical tests were considered significant when P<0.05. 88

PAGE 89

Results Localization of StAR Expression of StAR occurred in follicular cells of primary and secondary follicles, and in granulosa and theca cells of vesicular follicles. In corpora lutea, granulosa lutein cells stained positively for StAR and were the dominant cell type within corpora lutea. Theca lutein cells could not be distinguished in sections stained with StAR, even at high magnification (400x), although low numbers of these cells were detectable in H&E sections. Expression of StAR in corpora lutea occurred within the cytoplasm of granulosa lutein cells, and often exhibited regional differences within cells. Lutein cell nuclei were easily distinguishable due to their lack of staining. Additionally, for 35% (n=13 of 38) of corpora lutea stained for StAR, peripheral regions were more intensely stained than central regions, which often contained connective tissue and few or no luteal cells. This peripheral concentration of StAR expression was observed in female #2 (9% of examined corpora lutea), female #3 (25%), and female #4 (67%). Intra-Animal Differences There were a minimum of 16 primary follicles and eight secondary follicles observed from a section of the left ovary of female #1, and while a small number of follicles (n=8) exhibited minimal staining, most showed no positive staining, resulting in a mean staining score of 0.3 for this ovary (Table 5-1). In the sections of left ovary examined for female #2, one developing vesicular follicle with no positive staining was observed, while sections examined from the right ovary contained 22 primary follicles and two large vesicular follicles with light to moderate StAR expression (Fig. 5-1). The right ovary of female #2 showed greater StAR expression, even though the left ovary possessed more grossly visible follicles. The resulting mean follicular staining score for female #2 was 1.8. Throughout four ovarian sections examined for female #3, 19 vesicular follicles were identified (left ovary: 14, right ovary: 5), but StAR expression was 89

PAGE 90

minimal and only occurred within two follicles, resulting in a mean staining score of 0.05 for female #3. Vesicular follicles were observed within all ovarian sections examined for female #4 (left ovary n=7, right ovary n=8) and some developing vesicular follicles were also observed. The mean staining scores for follicles on the left (n=38) and right (n=27) ovaries of female #4 were 0.5 and 0.4, respectively, yielding an overall mean of 0.5 (n=65) for this female. There were no significant staining differences for follicles on the left and right ovaries of females #3 (P=0.384) and #4 (P=0.653). Corpora lutea examined for female #2 (n=19) exhibited StAR expression that ranged from low (n=6) to intense (n=3), yielding relatively high mean staining scores for the left (1.9, n=13) and right (1.8, n=6) ovaries (Table 5-2, Fig. 5-2). For female #3, aborting a mid gestational fetus, staining was absent from the left ovary (n=2), and a staining score of 0.5 was observed for the right ovary (n=2). The corpora lutea from female #4 (n=15) exhibited light (n=1), moderate (n=7), and intense (n=7) staining, with a staining score of 2.2 for the left ovary (n=7) and a score of 2.4 for the right ovary (n=8). There were no differences in stain intensity between the left and right ovaries of females #2 (P=0.961), #3 (P=0.667), and #4 (P=0.463). Inter-Animal Differences Expression of StAR within follicles was significantly greater for female #2 (early diestrus) than for the other three females, although all females exhibited follicles without StAR expression (Table 5-1). Expression of StAR within corpora lutea was significantly lower for female #3, who was undergoing abortion, than for females #2 (early diestrus) and #4 (late pregnancy) (Table 5-2, Fig. 5-3). 90

PAGE 91

Discussion Localization of StAR Immunohistochemical staining of manatee ovarian tissue with a polyclonal StAR antibody resulted in variable levels of staining in ovarian follicles and corpora lutea. As in Chapter 4, the presence of multiple follicles and corpora lutea within the ovaries of individual females helped to greatly increase sample size. Staining was confined to the cytoplasm of manatee luteal cells, as has been documented for human luteal cells [138], and intra-luteal variation in StAR expression detected at high magnification (1,000x) likely indicated the location of mitochondria. We hypothesized that developing or degenerating corpora lutea would exhibit regional differences in StAR expression, associated with such factors as the direction of luteal cell proliferation in developing corpora lutea (from the periphery, towards the center). However, this regional StAR expression was observed within corpora lutea at all stages of development, with greater StAR expression typically occurring in more peripheral areas. Such regional differences in StAR expression have also been identified in human corpora lutea, but with a reversed pattern, where the most intense staining occurs centrally. It has been suggested that differential stain intensity within a single human corpus luteum may be indicative of intraluteal regulation for expression of the StAR gene [138], which could explain this observance in the manatee. However, the greatest percentage of corpora lutea that exhibited concentrated peripheral StAR expression occurred in female #4 (late pregnancy), while the lowest percentage was observed for female #2 (early diestrus). This finding suggests that the trend towards peripheral StAR expression in the manatee may be related to corpus luteum age, where cells in the central regions of the corpus luteum regress first. While the staining intensity of granulosa and theca lutein cells could not be compared, the predominance of granulosa lutein cells suggests that these cells contribute more greatly to steroid 91

PAGE 92

production in the manatee corpus luteum, which is also typical for other species [136], except humans, where theca lutein cells express greater concentrations of StAR than granulosa lutein cells [138]. Intra-Animal Differences We hypothesized that fully functional follicles and corpora lutea would exhibit the most intense StAR expression, with lower expression indicating reduced steroid hormone production. Based upon morphological analyses (Chapter 4), it was thought that the steroidogenic potential of corpora lutea within individual females might also be variable, which was confirmed with StAR expression. These findings of differential function may be associated with variable age. Additionally, the lack of statistical difference in StAR expression between the left and right ovaries of females suggests that both ovaries contribute equally to steroid production during diestrus or pregnancy. During estrus, follicular expression of StAR would be associated with estrogen production [139], while an absence of staining would be an indication of follicular inactivity. However, a lack of expression in developing follicles would suggest the onset of atresia. The StAR expression observed in follicles during diestrus and pregnancy may implicate these structures for accessory corpora lutea development. Other species, including the mare [25], utilize accessory corpora lutea as a supplemental progesterone source to maintain pregnancy. Vesicular follicles have also been observed in pregnant African elephants during early gestation [62,105] and the dugong is reported to develop excess vesicular follicles [140]. Accessory corpora lutea formation in the manatee would account for vesicular follicle presence during diestrus and pregnancy, as well as the origin of some of this species numerous corpora lutea. Manatees are polyovular and typically monotocous [104,128], as are their relatives [105,140,141]. Based upon observed variation in StAR expression within female manatees, as 92

PAGE 93

well as morphological differences in these corpora lutea (see Chapter 4), it is proposed that differences in StAR expression among the corpora lutea of individual females could be the result of asynchronous ovulation and luteinization during the estrous cycle. Such asynchronous events could benefit the manatee, which ranges over a large habitat [142], is not highly social [143] (except when forced to aggregate due to cold weather [144]), and employs a promiscuous breeding strategy [145] with mating herds that can last for several weeks [1]. These factors may provide an added challenge to reproduction that would be mitigated by polyovularity and multiple, non-simultaneous ovulations that would increase the time within which a successful mating and implantation could occur. Multiple ovulations may also introduce the possibility of oocyte competition, just as sperm competition has been proposed for this species [146], or simply improve implantation rates by increasing the likelihood of oocyte fertilization. This finding challenges the hypothesis of Marmontel [104], who believed that all corpora lutea in a manatee ovary were formed from a single ovulation and luteinization event, but such a strategy would help explain differences in StAR expression among the corpora lutea of a diestrus female (#2) observed in this study. Despite minimal expression of StAR in some late pregnancy corpora lutea, female #4 still showed the most StAR expression and received the highest corpora lutea staining score (2.4) in the study. This finding indicates that manatee corpora lutea continue to produce progesterone throughout gestation, as this female was in very late pregnancy. The findings of a separate study (Chapter 4) indicated a reduction in corpus luteum and luteal cell size by late pregnancy in the manatee, but the intense expression of StAR in a late pregnancy female suggests that physical regression of luteal cells may precede functional regression in this species. 93

PAGE 94

Inter-Animal Differences We hypothesized that variation in StAR expression would be associated with different reproductive stages, and this was confirmed for both follicles and corpora lutea. Expression of StAR within follicles was significantly greater during early diestrus than anestrus and mid or late pregnancy, but expression of StAR was present at these other times. The occurrence of the most follicular StAR expression with female #2 (early diestrus) was expected because diestrus immediately follows follicular development and ovulation, and is a time when luteinization of unovulated follicles may occur. The observance of some StAR expression among the follicles of female #1 (anestrus) suggests that this female may have been preparing to enter another estrous cycle. The very low StAR expression among follicles of female #3 was probably related more to her impending abortion and death than to her pregnancy, because the other pregnant female (#4) did exhibit some follicular StAR expression. The absence of intense staining of follicles in this study is believed to be due to the fact that none of the specimens were in estrus, which would be the stage characterized by follicular growth and StAR expression associated with estrogen production. Expression of StAR within corpora lutea was significantly lower for female #3, who was undergoing abortion, than for females #2 (early diestrus) and #4 (late pregnancy). The relatively high corpora lutea staining score for female #2 was not unexpected, considering that in humans, granulosa lutein cell StAR expression is maximized in mid-diestrus [138]. It is likely that StAR expression in this female would have increased as luteal cell proliferation was completed. The near-absence of StAR expression within the corpora lutea of female #3 reinforces the hypothesis that this female was undergoing abortion. While the duration of luteal regression in the manatee is unknown, the 8-12 g bovine corpus luteum is fully reabsorbed within 4-6 days of the start of regression [86]. It appears that the two days which separated female #3s boat strike and death 94

PAGE 95

provided an adequate amount of time for the onset of luteal regression, based on the extremely low expression of StAR. In humans, StAR expression continues during regression, but its intensity is reduced [136]. If the same is true of the manatee, that would suggest a very advanced stage of regression for female #3 based on the near absence of StAR. Future Studies Further study of manatee corpora lutea should include experiments to extract steroid hormones from fresh luteal tissue and measure progesterone concentrations among corpora lutea of individual pregnant females, as well as from females at various stages of gestation. Also, serial sectioning of manatee corpora lutea in early pregnancy to search for a degenerating oocyte or zona pellucida could be used to confirm the presence of accessory corpora lutea. Conclusion Expression of StAR in manatee follicular cells and granulosa lutein cells was useful for determining the steroidogenic potential of follicles and corpora lutea. The greatest follicular expression of StAR was observed during early diestrus, while the greatest luteal StAR expression occurred in association with late pregnancy. The least StAR expression for both follicles and corpora lutea occurred in a female undergoing abortion. A commercial antibody, not specific to manatees, was used in this study, and as a result, the amount of epitope binding may have been reduced compared to what a manatee-specific antibody would achieve. However, in the absence of such an antibody, this commercial antibody was capable of showing relative differences in StAR expression within manatee follicles and corpora lutea. The application of StAR is particularly valuable for protected wildlife species such as the manatee, where the estrous cycle and pregnancy cannot be controlled or manipulated. In these cases, StAR can be used to try to determine the relative age of follicles and corpora lutea based on their steroidogenic potential, thus improving knowledge of the estrous cycle and pregnancy. 95

PAGE 96

Table 5-1. Follicle Summary Data ID (Status) # of Ovary Sections Examined No Stain (0) Minimal Stain (1) Moderate Stain (2) Intense Stain (3) Left Ovary Mean Stain Right Ovary Mean Stain Overall Mean Stain #1 (Anestrus) 1 n=16 n=8 --0.3 (n=24) -0.3 (n=24) #2 (Early Diestrus) 20 n=1 n=12 n=10 -0.0 (n=1) 1.9 (n=22) 1.8 (n=23) #3 (Mid Pregnancy) 4 n=17 n=2 --0.07 (n=14) 0.0 (n=5) 0.05 (n=19) #4 (Late Pregnancy) 15 n=35 n=25 n=5 -0.5 (n=38) 0.4 (n=27) 0.5 (n=65) Table 5-2. Corpora Lutea (CLs) Summary Data ID (Status) Total # of CLs # of CLs Examined No Stain (0) Min. Stain (1) Mod. Stain (2) Intense Stain (3) Left Ovary Mean Stain Right Ovary Mean Stain Overall Mean Stain #2 (Early Diest.) 34 19 -N=6 n=10 n=3 1.9 (n=13) 1.8 (n=6) 1.9 (n=19) #3 (Mid Preg.) 25 right* 4 n=3 N=1 --0.0 (n=2) 0.5 (n=2) 0.2 (n=4) #4 (Late Preg.) 56 left* 15 -N=1 n=7 n=7 2.2 (n=8) 2.4 (n=7) 2.4 (n=15) Pregnant uterine horn: left or right. A B Figure 5-1. A primary follicle from a female in early diestrus stained with StAR (A) and H&E (B). The section stained with StAR exhibits moderate steroidogenic activity. Numerous macrophages (arrows) are visible in both A and B, indicating that this is likely a dying follicle. (400x) 96

PAGE 97

A B C D Figure 5-2. Differential StAR staining in the corpora lutea of female #2. The section in A is moderately stained, B is moderately to intensely stained, and C is intensely stained. Arrows denote a positively stained granulosa lutein cell. D is an H&E section of a corpus luteum from female #2, containing numerous granulosa lutein cells. (250x) 97

PAGE 98

A1 A2 B1 B2 Figure 5-3. StAR and corresponding H&E images for 2 pregnant female manatees. The staining in female #4 (late pregnancy; B1 & B2) is more intense than in female #3 (mid pregnancy; A1 & A2). Connective tissue (arrows) is visible among luteal cells in A2 (female #3) and B2 (female #4), but is more prominent in A2 (#3). (x250) 98

PAGE 99

CHAPTER 6 EVALUATION OF PERACUTE AND CHRONIC STRESS IN THE FLORIDA MANATEE Introduction The Florida manatee lives in a challenging environment with threats that include watercraft collision, line entanglement, red tide, and cold stress. There has been a significant research effort to promote manatee conservation, which includes brief capture of free-ranging manatees for health assessments as well as rescue of injured or sick individuals. However, investigations of manatee stress (both peracute and chronic) that may be associated with capture, injury, or disease have been limited [26]. This is due in part to the short duration of health assessments that are now typically completed within one hours time, after which the animals are released. Thus, the current handling procedure for this species limits opportunities to collect data on stress and stress indicators. Among animals, ACTH (adrenocorticotropic hormone) is secreted from the anterior pituitary within minutes of peracute stress exposure and triggers glucocorticoid secretion from the adrenal cortex [27]. The glucocorticoid found in birds and small mammals is corticosterone, while cortisol (hydrocortisone) occurs in medium to large mammals, as well as fish and humans [28]. Circulating ACTH peaks and subsides quickly, while a glucocorticoid peak with a 30-60 minute duration [27] occurs 1 to 4 hours after encountering the peracute stressor [29-34]. The intensity of the stressor determines the degree and duration of the endocrine stress response [27,30]. During the stress response, glucocorticoids help provide energy by upregulating lipolysis and glucose formation, altering metabolic pathways and vasculature via catecholamines, and generally decreasing immune response in order to minimize cell and tissue damage [35,36]. However, long periods of heightened glucocorticoid concentrations are detrimental as they shift energy away from other important biological processes. Problems associated with prolonged 99

PAGE 100

glucocorticoid increase include disruption of the circadian rhythm, metabolism, growth, wound healing, and gastrointestinal function, as well as immunosuppression, neuronal cell death, hyperglycemia, hypertension, insulin resistance, muscle and bone atrophy, inhibition of reproductive cycles, abortion, and death [27,29,37]. There is disagreement over the level and duration of increased serum cortisol concentration that can be harmful, but most studies agree that more than several days of stress can begin to cause physiological effects [28,30,38,39]. The ability to differentiate between physiologic and pathologic glucocorticoid concentrations is critical, because glucocorticoids have many functions that vary by species, gender, age, and reproductive status [28,147]. In females, for example, glucocorticoids naturally fluctuate throughout the estrous cycle and pregnancy [29]. Glucocorticoids also exhibit daily and seasonal variation [30,147] that could interfere with result interpretation. Similarly, natural fluctuations in ACTH concentrations must be distinguished from those associated with varying degrees of peracute stress. Previous peer-reviewed studies of manatee stress have measured cortisol [148,149]. The first study [148] measured cortisol in 20 free-ranging, captured female and male manatees and reported a mean cortisol concentration of 0.1 g/dl. The later study, involving two long-term captive males [149], reported a mean cortisol concentration of 0.2 0.1 g/dl. These manatees were also subjected to a transport simulation in which they were removed from their habitat, kept moist on foam pads for 6 hours, and subjected to some restraint. A subsequent cortisol concentration of 0.8 g/dl was reported for one of these manatees, while a concentration of 2.1 g/dl was observed in the second manatee, within 4.5 hours and 7.5 hours of the start of the transport, respectively (C. Manire pers. comm.). 100

PAGE 101

The objectives of this study were to: (1) determine the predominant glucocorticoid in the manatee; (2) analytically validate chemiluminescent assays to assess the predominant glucocorticoid and ACTH within manatee serum or plasma; (3) determine diagnostic thresholds to distinguish physiological concentrations of stress indicators from pathological concentrations; and (4) identify the most suitable methods for diagnosing peracute and chronic stress for manatees, given the current handling protocol for this endangered marine mammal. Materials and Methods Blood Sampling Blood samples were opportunistically collected from the brachial vascular bundle during captures of free-ranging manatees, as part of ongoing, government-led ecological assessments [46,47], or during health assessments of captive manatees at facilities throughout Florida. Serum and lithium heparin plasma (LiHep) for cortisol analysis and EDTA plasma for ACTH analysis were separated by onsite centrifugation (3,000 x g for 10 min.), typically within 1 hour of collection, and never more than 4 hours after collection as recommended in the literature [150]. Samples aliquotted for cortisol analyses were then transported on ice and either assayed immediately upon arrival at the University of Florida or stored at -80C and analyzed at a later date. EDTA plasma aliquots were frozen immediately after centrifugation using either dry ice or liquid nitrogen, in which they were stored during transport to the University of Florida. Samples were then stored at -80C until analysis, which was within 4 months of the collection date. When thawed for assay, ACTH samples were maintained at or below 4C until analysis. Sample Population A total of 127 cortisol and 53 ACTH samples were analyzed. The cortisol sample population consisted of females (n=84) and males (n=43), and included calves (n=15), subadults (n=21), and adults (n=91). For ACTH, female (n=29) and male (n=24) samples were tested, 101

PAGE 102

including calves (n=6), subadults (n=6), and adults (n=41). Calves were defined by a total straight length of 82 to 245 cm, subadults were 246 to 265 cm, and adults were >265 cm [50]. Apparently healthy female and male subadult and adult manatees were divided into stress categories based on origin (captive, free-ranging, or rehabilitating); serum amyloid A (SAA) concentration, where concentrations >70 g/ml are associated with inflammatory disease [53]; and lactate, where concentrations >20 mmol/L indicate moderate exertion [151]. These criteria were used to establish three stress-related categories in the sample population: unstressed, peracutely stressed, and chronically stressed. Unstressed manatees were those in captivity for multiple years (including healthy, rehabilitated manatees) who were accustomed to handling for blood draw, with SAA <70 g/ml and lactate <20 mmol/L. Peracutely stressed manatees were free-ranging individuals captured for health assessment or manatees recently transported to a rehabilitation facility, which represented a potent stimulus experienced in the past 0-4 hours. To be included in the peracutely stressed category, these manatees needed SAA <70 g/ml to eliminate inclusion of animals with chronic inflammatory disease, but lactate could be >20 mmol/L. Chronically stressed manatees were typically those in a rehabilitation setting undergoing medical treatment for an injury or other condition that persisted for more than several weeks, but also included some long-term captive manatees. These chronically stressed individuals were further characterized by SAA < or >70 g/ml and lactate <20 mmol/L. Serum cortisol and ACTH ROC analyses included samples collected in all seasons of the year. The unstressed (n=20) category consisted of 12 females (n=3 subadults, n=9 adults) and eight males (n=4 subadults, n=4 adults). There were 60 peracutely stressed manatees in the study population, which was comprised of 34 females (n=6 subadults, n=28 adults) and 26 males (n=4 subadults, n=22 adults). Finally, the chronically stressed (n=31) population consisted of 28 102

PAGE 103

females (n=3 subadults, n=25 adults) and three males (n=1 subadult, n=2 adults). For analysis of plasma ACTH concentration, there were six unstressed manatees (captive adults: n=4 female, n=2 male). There were 32 peracutely stressed manatees included in the ACTH ROC analysis, including 13 females (n=2 subadults, n=11 adults) and 19 males (n=4 subadults, n=15 adults), of which 27 were free-ranging and five were captive. The chronically stressed group consisted of nine captive female manatees (n=9 adults). To establish baseline cortisol and ACTH concentrations for the manatee, samples were collected and assayed from two adult, male, healthy, long-term captive manatees that are trained to submit to a multitude of handling procedures, including blood draw (manatee 1 n=9, manatee 2 n=2). Samples from these two males as well as from other unstressed subadult and adult manatees in the sample population were examined with respect to season for cortisol (n=20 manatees; n=24 samples: winter n=8, spring n=2, summer n=11, autumn n=3) and ACTH (n=6 manatees; n=8 samples: winter n=3, spring n=2, summer n=1, autumn n=2). ACTH results from the sample population were also compared to capture duration (n=30), lactate concentration (n=42) measured from standard biochemical analysis of lithium heparin plasma (University of Florida College of Veterinary Medicine, Department of Clinical Pathology, Gainesville, FL), and respiratory rate per 5 minute interval (n=20). Differences in sample size were due to sample collection and assessment logistics in the field capture setting. Laboratory Analysis Manatee serum samples (n=12) were submitted to the Animal Health Diagnostic Center at Cornell University (Ithaca, NY) for measurement of corticosterone via a serum radioimmunoassay (MP Biomedicals, Solon, Ohio) to determine if this glucocorticoid was present in the manatee. An automated chemiluminescent immunoassay analyzer was used to assay cortisol and ACTH (IMMULITE 1000, Siemens Medical Solutions Diagnostics, Los 103

PAGE 104

Angeles, CA) using polyclonal rabbit anti-cortisol (competitive assay; product no. LKC01, Siemens Medical Solutions Diagnostics, Los Angeles, CA) and monoclonal murine anti-ACTH antibodies (sandwich assay; product no. LKAC1, Siemens Medical Solutions Diagnostics, Los Angeles, CA), respectively. The cortisol assay used a single reagent (alkaline phosphatase conjugated to cortisol in buffer, with preservative) while the ACTH assay utilized two reagents (a protein buffer/serum matrix and an alkaline phosphatase conjugated to polyclonal rabbit anti-ACTH antibody in buffer, with preservative). Sample volumes were per the manufacturers specification sheet. Analytical validation of this system using manatee serum and LiHep plasma (for cortisol) and EDTA plasma (for ACTH) was completed at the University of Floridas endocrine laboratory at the College of Veterinary Medicine to ensure cross reactivity of antibodies with manatee serum and/or plasma. An attempt was also made to measure ACTH using LiHep plasma, but ACTH concentrations observed with LiHep were not sufficiently correlated (60%, n=12) with EDTA plasma results [60], indicating that EDTA plasma, recommended by the manufacturer, is required for this assay. To analytically validate this assay, experiments were completed to determine precision (intra-assay and inter-assay coefficients of variation), accuracy, analytical sensitivity, and methods similarity. Analytical specificity was not tested as part of this study, but was established by the manufacturer for both cortisol (Siemens PILKCO-8, 2004-09-09) and ACTH (Siemens PILKAC-12, 2006-12-29). For all analytical validation experiments, samples with low, intermediate, and high concentrations of analyte were tested. Intra-assay precision was tested using 10 simultaneous replicate samples from three manatees for serum and lithium heparin cortisol and ACTH, while inter-assay precision was tested using three replicate samples from three different manatees for each analyte. Accuracy was assessed by dilution experiments 104

PAGE 105

to test analyte parallelism and recoverability using samples from three manatees for cortisol (serum and LiHep) and ACTH. Assay analytical sensitivity, or the lower limit of detection (LLOD), which can be affected by protein and plasmatic non-specific interference, was determined by assaying ten replicates of the zero standard (cortisol-free human serum, with preservative or ACTH-free bovine protein-based matrix, with preservative). For methods comparison, the IMMULITE system at UF was compared to a radioimmunoassay (Siemens coat-a-count) (serum and LiHep cortisol; n=10) and another IMMULITE 1000 system (Siemens) (serum and LiHep cortisol, ACTH; n=10) at Cornell Universitys Animal Health Diagnostic Center using paired samples. Statistical Analysis Validation results (precision, accuracy, methods comparison, and ROC) were calculated with EP Evaluator Release 7 (David G. Rhoads Associates, Inc., Kennett Square, PA). Precision was assessed by coefficients of variation (CV). Allowable error was defined (25% and 0.2 g/dl) for determination of accuracy, which was then assessed by linearity. Linearity was achieved when the dilution results did not differ from the expected results by a percentage greater than the systematic allowable error (50% of the total allowable error budget). Accuracy was further evaluated using adjusted-R 2 values resulting from linear regression, where results closer to 1.0 indicated a better relation between the independent and dependent variables. Methods were compared using Deming regression [56] and evaluated using total allowable error (TEa) [57] (<25%) and the observed correlation coefficient (r). Receiver operating characteristic (ROC) analyses [51] were used to determine cortisol and ACTH concentration ranges associated with different stress categories for diagnostic validation. For each ROC analysis, diagnostic sensitivity (tests probability of producing a true positive result) and specificity (tests probability of producing a true negative result), and positive and 105

PAGE 106

negative predictive values were determined. Positive predictive value (PPV) is the probability that a positive test result accompanies a true positive condition [# True Positives / (# True Positives + # False Positives)] while negative predictive value (NPV) is the probability that a negative test result accompanies a true negative condition [# True Negatives / (# True Negatives + # False Negatives]). Sigma Stat (Systat Software, Inc., Point Richmond, CA) was used to complete all other statistical tests. To determine gender differences, balanced subpopulations of males and females were created using equal numbers of samples from each gender, age class, and season. For cortisol, this resulted in a population of 31 females and 31 males, which were compared using a one-way ANOVA on ranks. For ACTH, 15 female and 15 male samples were tested using a one-way ANOVA. To test differences in cortisol (unstressed n=20; peracutely stressed n=60; chronically stressed n=31) and ACTH (unstressed n=6; peracutely stressed n=32; chronically stressed n=9) among stress categories, one-way ANOVAs on ranks (Dunns method) were used. A t-test (Mann-Whitney Rank Sum Test) was used to compare cortisol concentrations from unstressed manatees for two portions of the year: summer and autumn (n=14) vs. winter and spring (n=10). Pearson correlations were used to test for correlation between ACTH and capture time (n=26) and lactate concentration (n=46) for subadult and adult manatees. A t-test was used to test for significant differences in ACTH concentration 20 minutes after capture for manatees with respiratory rates 5 breaths (n=8) or >5 breaths / 5 minute interval (n=12). Statistical tests were determined to be significant when P<0.05. 106

PAGE 107

Results Glucocorticoid Determination For 11 of the 12 samples tested for corticosterone, the result was below the LLOD of the assay (7.7 ng/ml) and the only result above the LLOD was 9.8 ng/ml, which represents a concentration lower than trace (2x LLOD). Assay Analytical and Diagnostic Validation Precision analyses resulted in coefficients of variation (CV) ranging from 9.1% to 16.7% for serum cortisol, 7.8% to 21.7% for LiHep cortisol, and 3.0% to 8.5% for ACTH. All accuracy results for cortisol (serum and LiHep) and ACTH were linear. The adjusted-R 2 values for accuracy testing were 0.87 for serum cortisol, 0.79 for LiHep cortisol, and 0.96 for ACTH. The assays analytical sensitivity was determined for manatee serum and LiHep cortisol (0.1 g/dl), as well as manatee EDTA plasma ACTH (10.0 pg/ml). The methods comparison for serum cortisol produced a Deming regression slope of 0.94 [95% CI: 0.76-1.13], an intercept of -0.04 [95% CI: -0.18-0.09], and a correlation coefficient (r) of 0.97. The LiHep cortisol comparison resulted in a slope of 0.82 [95% CI: 0.60-1.03], an intercept of 0.07 [95% CI: -0.09-0.23], and r=0.95. The slope for ACTH methods comparison was 0.91 [95% CI: 0.77-1.04], with an intercept of 8.92 [95% CI: -17.90-35.73], and r=0.98 (Table 6-1). These results demonstrated very similar results between the methods. Cortisol and ACTH data were collected from manatee calves but were not included in ROC analyses or other tests. These calf data are presented in Table 6-2, along with summary data for subadults and adults. There were no gender differences in cortisol (P=0.504) or ACTH (P=0.984) among subadults and adults, so genders were grouped for ROC analysis. ROC analyses of pairwise comparisons of cortisol concentrations among the three stress categories indicated thresholds at which sensitivity, specificity, and positive and negative 107

PAGE 108

predictive values were optimized. A cortisol threshold of 0.2 g/dl distinguished unstressed and peracutely stressed manatees, while 1.1 g/dl best differentiated peracutely and chronically stressed manatees, and a threshold of 0.5 g/dl differentiated unstressed and chronically stressed manatees (Table 6-3). However, when specificity alone was maximized (100%) for each comparison (to minimize false positive results), thresholds near 1.0 g/dl were most diagnostic for pairwise comparisons of unstressed vs. peracutely stressed (0.9 g/dl), peracutely vs. chronically stressed (1.3 g/dl), and unstressed vs. chronically stressed (1.0 g/dl) manatees. In these cases where specificity was maximized, the positive predictive value (PPV) was also reported as 100%. Additionally, cortisol concentrations in chronically stressed manatees were significantly greater than concentrations in unstressed and peracutely stressed individuals (Table 6-4) and cortisol concentrations in chronically stressed manatees were up to 53 times greater than those observed in some unstressed females (5.3 vs. 0.1 g/dl). Pairwise comparisons by stress category were also completed for ACTH results. When all parameters were optimized, a threshold of 67.3 pg/ml differentiated unstressed and peracutely stressed manatees, while 87.5 pg/ml best differentiated peracutely and chronically stressed manatees, and a threshold of 55.5 pg/ml differentiated unstressed and chronically stressed manatees (Table 6-5). When only specificity and PPV were optimized (100%), slightly different thresholds were found to be diagnostic for unstressed vs. peracutely stressed (87.5 pg/ml), peracutely vs. chronically stressed (199 pg/ml), and unstressed vs. chronically stressed (82.5 pg/ml) manatees. Peracutely stressed manatees had significantly higher ACTH concentrations than unstressed or chronically stressed individuals, while there was no significant difference in ACTH concentration between unstressed and chronically stressed manatees (Table 6-4). 108

PAGE 109

Population Baseline and Seasonal Variation Baseline cortisol (0.2 g/dl) and ACTH (16.4-58.4 pg/ml) concentrations for the manatee were determined. When cortisol and ACTH concentrations from unstressed manatees were examined with regard to seasonal differences, no seasonal cortisol difference was observed for summer and autumn vs. winter and spring (P=0.681). While low sample size prevented statistical analysis of seasonal ACTH results, mean results and ranges were similar for all seasons except summer, for which only a single sample with low ACTH concentration was available (Table 6-6). Correlation of ACTH with Capture Parameters Total capture time was positively correlated with ACTH concentration (r=0.467, P=0.0162). The maximum ACTH concentration observed in the current study (643 pg/ml) occurred in a healthy manatee followed for 50 minutes prior to successful capture, which represents a greater than normal follow time. Lower values occurred in healthy manatees captured in a shorter time (67 pg/ml), but even the ACTH concentration associated with a follow time of 15 minutes (145 pg/ml) was significantly greater than that of the unstressed population (mean +1.96SD; 78.9 pg/ml) (Table 6-7). Lactate and ACTH were also positively correlated (r=0.573, P=0.0000720). Concentrations of ACTH were not significantly different in manatees sampled 20 minutes after capture, with respiratory rates 5 breaths/minute or > 5 breaths/minute (P=0.847). Concentrations of ACTH and lactate and capture time showed no discernable relationship to serum cortisol concentrations (Table 6-7). Discussion Glucocorticoid Determination It was hypothesized that hydrocortisone (cortisol) would be the predominant manatee glucocorticoid, as it is in other large mammals and marine mammals that have been studied [31, 109

PAGE 110

152-154]. Our hypothesis was confirmed by the observation of trace corticosterone concentrations in all 12 samples. It is possible that the corticosterone antibody utilized in the present study did not cross-react with corticosterone in manatee plasma, but this is unlikely due to the stable structure of this molecule. Furthermore, in oldfield mice, a species for which corticosterone predominates, baseline levels reach 523 ng/ml and exceed 1,900 ng/ml in stressed individuals [155]. Assay Analytical and Diagnostic Validation It was hypothesized that commercially available chemiluminescent assays could be validated for measuring cortisol and ACTH in manatee serum and plasma. The manufacturer has not validated the cortisol assay for use with LiHep plasma, but there are no data to suggest that plasma should not be utilized for this assay (Siemens Technical Support pers. comm.). Both the cortisol and ACTH assays were analytically validated, with precision metrics within allowable standards (CV <25% for cortisol immunoassays and <10% for ACTH) [58], linear accuracy results [59], and correlation coefficients >0.97 for serum cortisol and ACTH methods comparison [60]. For cortisol, ROC analyses were completed using only serum results because all available plasma data represented manatees for which serum cortisol results had been obtained. Calf data were excluded from statistical analyses due to inherent differences in cortisol concentration known to exist in neonates and young of other species [27]. It was further hypothesized that unstressed, peracutely stressed, and chronically stressed manatees could be differentiated based on observed cortisol and ACTH concentrations, and that such categorization would define physiologic and pathologic concentrations of manatee cortisol and ACTH. The observation of similar cortisol concentrations in peracutely stressed (mean= 0.5 g/dl) and unstressed manatees (mean=0.4 g/dl) indicated that cortisol cannot be used as a peracute stress indicator for this species. This was further supported by ACTH values that 110

PAGE 111

increased as peracute stress increased, without a similar graded increase in cortisol concentration (Table 6-7). These findings suggest that manatees are like other animals whose cortisol does not peak until at least 1 to 4 hours after exposure to a peracute stressor [29,31-34]. Because manatees are typically released within 1 hour of their capture, this finding is believed to be due, in part, to the fact that blood samples are collected prior to the physiological cortisol increase associated with peracute stress. The findings of a previous study [149] indicate that cortisol does increase in response to peracute stress in the manatee, as two manatees exhibited a peak in cortisol concentration just prior to the end of a transport simulation lasting 4.5 and 7.5 hours, respectively. While the use of cortisol for peracute stress diagnosis is further complicated by its numerous physiologic roles [29], cortisol may be used to assess chronic stress that results from disease or environmental factors including cold stress syndrome [156]. Analyses of unstressed and chronically stressed manatees indicated that cortisol concentrations 1.0 g/dl are diagnostic of chronic stress (100% specificity and PPV). In an African and Asian elephant, cortisol concentrations were found to increase 4-8 fold over a 0.5 g/dl baseline concentration, 30 minutes after an injection of ACTH [157]. These concentrations of 2-4 g/dl were similar to cortisol concentrations observed in chronically stressed manatees, which ranged from 0.1-5.3 g/dl, with a mean of 1.0 g/dl. If elephants were administered a second ACTH injection (equivalent to exposure to a second peracute stressor), cortisol concentrations of 4.5-6.0 g/dl were observed [157], which are approximate to the maximum cortisol value observed in the manatee (5.3 g/dl). Similar results occurred following ACTH stimulation in four adult male Asian elephants, with serum cortisol concentrations 3-10 fold above baseline [158]. 111

PAGE 112

This paper documents manatee ACTH, which can be used as an indicator of peracute stress in this species for which a fight or flight stress response has been demonstrated. Concentrations of ACTH were observed to increase rapidly and in proportion to the capture or handling stress experienced, with concentrations 87.5 pg/ml diagnostic of peracute stress. This assay requires EDTA plasma, which is not typically currently banked during manatee captures and health assessments. Therefore, if there is a continued interest in assessing manatee ACTH, collection of EDTA plasma must be implemented. Also, because ACTH is such a small protein hormone, plasma must be centrifuged and frozen immediately after collection to prevent ACTH degradation. Population Baseline and Seasonal Variation It was hypothesized that two captive males previously assessed for cortisol [149] could be included in the current study to establish baseline cortisol and ACTH concentrations for the manatee. Using these two males, a mean and median cortisol concentration of 0.1 g/dl was observed, which is believed to represent the baseline cortisol value for this species. This value is also similar to published cortisol concentrations for captive female elephants (mean = 0.06 0.01 g/dl [159]). The manatees cortisol baseline was further confirmed by the ROC threshold of <0.2 g/dl found to be diagnostic of unstressed captive and rehabilitated manatees. Baseline manatee ACTH concentrations (16.4-58.4 pg/ml) were also obtained from the two long-term captive males. While these two males exhibited low cortisol and ACTH concentrations, such values may not be exhibited by all healthy captives due to possible aggression in the absence of mating, or other social issues that may increase stress indicators. Seasonal assessments of cortisol and ACTH were made using only unstressed individuals in order to eliminate confounding factors associated with peracute and chronic stress. Cortisol data were binned into two seasonal time frames for statistical analysis, but low sample sizes 112

PAGE 113

prevented statistical analysis of ACTH data. While low sample size prevented complete statistical analysis of seasonal data, the observed findings suggest that there are not strong physiologic seasonal differences in manatee cortisol and ACTH concentrations. Correlation of ACTH with Capture Parameters We hypothesized that ACTH would be positively correlated with capture time and lactate concentration and that there would be significant differences in ACTH concentration associated with respiration rate in free-ranging manatees, because we perceived these to be parameters affected by peracute stress. However, the only significant correlations occurred with capture time and lactate. While lactate is constantly produced during exercise, increased lactate concentrations occur when lactate is produced more rapidly than it can be removed from tissues during intense physical exercise. As manatees evade capture for longer periods of time, lactate accumulates due to prolonged exertion, and ACTH increases because of peracute stress associated with this fight or flight response. Higher median lactate concentrations (4x) in free-ranging versus captive manatees were also observed in another study [160]. Distinct peracute stress was observed among many captured free-ranging manatees, particularly a female pursued for 50 minutes prior to successful capture (ACTH=643 pg/ml). While pursuits of this duration are uncharacteristic of the manatee research program, the data observed in this study support documenting and minimizing pursuit, as even follows as short as 15 to 20 minutes in duration resulted in ACTH concentrations significantly greater than those of unstressed individuals (Table 6-7). In lieu of ACTH data, capture time and lactate concentration could be used to help assess stress, as increases in these parameters would be associated with increased peracute stress. The lack of significant difference in ACTH concentration between manatees with normal (5 breaths / 5 minutes) versus elevated respiratory rates (>5 breaths / 5 minutes) 20 minutes after capture 113

PAGE 114

may have been due to the small sample size available for testing or individual variation in respiratory patterns and stress response. Future Studies A separate study to investigate the manatees molecular stress response (MSR) is currently underway. The MSR is activated at the cellular level of all organisms within hours of stressor exposure to restore homeostasis [161-164]. The MSR continues until homeostasis is regained [164] and stimulates such processes as oxidative stress repair and cellular detoxification, and the regulation of cell growth and death. When the MSR remains activated for prolonged periods, it can lead to disease and accelerated aging [165]. Immunohistochemical staining of manatee skin samples is being used to detect 40 different stress response proteins associated with the MSR. This methodology has also been applied to elephants [165] and may prove to be a preferable method for quantifying long-term manatee stress associated with disease or environmental conditions. A future study should also complete the evaluation of seasonal variation of cortisol and ACTH in unstressed individuals that was begun in the present study. Diurnal variation should also be examined, and was attempted in the present study, but could not be completed due to time constraints. Such studies are important because glucocorticoids and ACTH exhibit a circadian rhythm in most species, wherein concentrations are low while an animal is active, and rise (5-10x) during sleep so that concentrations are highest in the morning, upon waking [30,147]. Such time-dependent variation could affect cortisol and ACTH concentrations observed in the blood. Conclusion In this paper, hydrocortisone (cortisol) was determined to be the predominant glucocorticoid in the manatee and chemiluminescent assays for cortisol and ACTH were 114

PAGE 115

analytically and diagnostically validated for manatee serum or plasma. Cortisol was diagnostic for chronically stressed manatees (1.0 g/dl), while ACTH was diagnostic for peracute stress (87.5 pg/ml). Capture time and lactate concentration were positively correlated with ACTH concentration. Cortisol and ACTH concentrations associated with normal physiologic variation, capture, and disease were defined, which will provide manatee biologists and clinicians with an additional tool to evaluate stress, given the manatees capture and handling protocol. Table 6-1. Cortisol and ACTH Analytical Validation Results Validation Metric Serum Cortisol LiHep Plasma Cortisol EDTA Plasma* ACTH Inter-assay Precision CV (sample min-max tested) 9.1 16.7% (0.6 1.1 g/dl) 7.8 21.7% (0.4 0.5 g/dl) 3.0 8.5% (46 327 pg/ml) Intra-assay Precision CV (sample min-max tested) 12.5 23.5% (0.6 1.5 g/dl) 9.3 19.6% (0.5 1.2 g/dl) 4.1 6.9% (99 400 pg/ml) Accuracy Adjusted-R2 (sample min-max tested) 0.87 (0.5 1.3 g/dl) 0.79 (0.4 1.0 g/dl) 0.96 (42 615 pg/ml) Methods Comparison correlation coefficient (r) 0.97 (97%) 0.95 (95%) 0.98 (98%) Methods Comparison Deming Regression Slope (95% CI) 0.94 (0.76 1.13) 0.82 (0.60 1.03) 0.91 (0.77-1.04) Methods Comparison Intercept (95% CI) -0.04 (-0.18 0.09) 0.07 (-0.09 0.23) 8.92 (-17.9-35.73) *EDTA plasma is required for ACTH analysis. LiHep plasma results correlate poorly with EDTA plasma results, making LiHep plasma unsuitable for this assay. Table 6-2. Cortisol and ACTH Results Associated with Different Age Classes Age Classes Cortisol Mean SD, Min-Max (n) ACTH Mean SD, Min-Max (n) Calves (82-245 cm) 0.5 0.3, 0.2-1.4 (15) 160.7 95.9, 82.5-341.0 (6) Subadults (246-265 cm) 0.6 0.7, 0.2-3.6 (21) 202.7 72.5, 123.0-283.0 (6) Adults (>265 cm) 0.6 0.6, 0.1-5.3 (90) 182.6 148.4, 26.7-643.0 (41) 115

PAGE 116

Table 6-3. Serum Cortisol Pairwise Stress Category ComparisonsThresholds Where All ROC Parameters Optimized ROC Parameter Unstressed vs. Peracutely Stressed Peracutely Stressed vs. Chronically Stressed Unstressed vs. Chronically Stressed Diagnostic Threshold 0.2 g/dl 1.1 g/dl 0.5 g/dl Sensitivity 100.0% 32.3% 90.3% Specificity 30.0% 96.7% 65.0% PPV 81.1% 83.3% 80.0% NPV 100.0% 73.4% 81.3% Table 6-4. Cortisol and ACTH Descriptive Statistics Among Three Stress Categories Stress Parameter Unstressed Peracutely Stressed Chronically Stressed Cortisol Mean SD (g/dl) 0.4 0.2 a (n=20) 0.5 0.2 a (n=60) 1.0 1.0 b (n=31) Cortisol Min-Max (g/dl) 0.1 0.8 0.2 1.2 0.1 5.3 ACTH Mean SD (pg/ml) 47.7 15.9 c (n=6) 240.5 136.7 d (n=32) 80.0 47.3 c (n=9) ACTH Min-Max (pg/ml) 26.7-75.7 67.3-643.0 32.6-197.0 Cortisol concentrations were significantly lower in unstressed a and peracutely stressed a manatees than in chronically stressed manatees b ACTH concentrations were significantly lower in unstressed c and chronically stressed c manatees than in peracutely stressed manatees d Table 6-5. Pairwise Stress Category Comparisons for ACTHThresholds Where All ROC Parameters Optimized ROC Parameter Unstressed vs. Peracutely Stressed Peracutely Stressed vs. Chronically Stressed Unstressed vs. Chronically Stressed Diagnostic Threshold 67.3 pg/ml <87.5 pg/ml 55.5 pg/ml Sensitivity 100.0% 96.9% 77.8% Specificity 83.3% 88.9% 83.3% PPV 97.0% 96.9% 87.5% NPV 100.0% 88.9% 71.4% Table 6-6. Seasonal Analyte Concentrations in Unstressed Manatees Analyte Winter Mean SD, Min-Max (n) Spring Mean SD, Min-Max (n) Summer Mean SD, Min-Max (n) Autumn Mean SD, Min-Max (n) Cortisol 0.4 0.2, 0.1 0.7 (8) 0.2 0.1, 0.1 0.3 (2) 0.4 0.3, 0.1 0.8 (11) 0.1 0.0 (3) ACTH 57.4 15.9, 47.1 75.7 (3) 40.0 7.6, 34.7 45.4 (2) 18.6 (1) 34.4 10.7, 26.8 42.0 (2) 116

PAGE 117

Table 6-7. Parameters Associated with Increasing Follow Time in Three Manatees ID Chase Time (min) Lactate (mmol/L) ACTH (pg/ml) COR (g/dl) TEP13 15 17 145 0.7 TEP12 20 18 191 1.2 TEP14 50 23 643 0.8 117

PAGE 118

CHAPTER 7 CONCLUSIONS The IMMULITE 1000s progesterone assay has proven to be a valid and highly sensitive and specific tool for diagnosing pregnancy in the manatee, particularly during the first 6 months of gestation. The assay will allow manatee biologists and veterinarians to monitor reproductive health in individual free-ranging and captive females and will be beneficial in managing captive breeding of Sirenians. Additional studies of manatee progesterone and pregnancy should monitor concentrations of this hormone throughout gestation, which would require sample collection from well-trained, healthy, captive or rehabilitating females, to minimize stress associated with handling. Analysis of serum progesterone throughout non-pregnant estrous cycles is also required to determine the duration and range of progesterone concentrations that characterize non-pregnant diestrus. Such an investigation would require the use of captive females housed with males to help guarantee normal reproductive cyclicity. The relatively low circulating P4 concentrations in the manatee cannot be attributed to metabolism, because P4 itself was most prominent within the plasma of females tested with GC/MS/MS. Several metabolites were detected, but the two most common were 5-P3-OH and 5-DHP. These metabolites were most prevalent during diestrus and early pregnancy, when maximum P4 concentrations were also observed. The decline of all circulating progestins in late pregnancy suggests that the placenta does not contribute to progestin production in the manatee. An investigation of progesterone binding proteins in manatee blood could help explain if the pattern of progesterone increase in early pregnancy is related to a decreased metabolic clearance rate for progesterone, as has been observed for the guinea pig and coypu [81,82]. Additionally, further study of the counter-current exchange system in the vasculature surrounding the female 118

PAGE 119

reproductive system of the manatee may demonstrate a localized route for progestin transport from the ovaries to the uterus, and help explain the low concentrations in peripheral circulation. The study of manatee corpus luteum morphology provided a general description of its structure, identification of its luteal cell types, documentation of fibroblast and endothelial cell morphology, and the first transmission electron micrographs of manatee luteal cells. For the first time, intra-female and inter-female differences in corpus luteum and luteal cell morphology have been documented, and related to changes in corpus luteum function. These findings provide new information about this species poorly understood reproductive trends and represent progress towards understanding the role of the manatees multiple corpora lutea. Further insight into manatee corpus luteum function was also found through the study of StAR. Expression of StAR in manatee follicular cells and granulosa lutein cells was useful for determining the steroidogenic potential of follicles and corpora lutea. Additionally, when StAR results were compared to morphometry findings, it was found that morphological regression of manatee corpora lutea may precede functional regression as corpora lutea which appeared to be degenerating were found to possess intense StAR expression. In the absence of an antibody specific for manatees, this commercially-available antibody was capable of showing relative differences in StAR expression within manatee corpora lutea. Further study of manatee corpora lutea should include experiments to extract steroid hormones from fresh luteal tissue and measure progesterone concentration within corpora lutea from the same pregnant female, as well as from females at various stages of gestation. Also, serial sectioning of manatee corpora lutea in early pregnancy to search for a degenerating oocyte or zona pellucida could be used to confirm the presence of accessory corpora lutea. 119

PAGE 120

The analysis of manatee stress indicators identified hydrocortisone (cortisol) as the predominant glucocorticoid in the manatee. Additionally, chemiluminescent assays for cortisol and ACTH were analytically and diagnostically validated to quantify these substances in manatee serum or plasma. Cortisol was diagnostic for chronically stressed manatees (1.0 g/dl), while ACTH was diagnostic for peracute stress (87.5 pg/ml). Cortisol and ACTH concentrations associated with normal physiologic variation, capture, and disease were identified, which will be useful to manatee biologists and clinicians and provide an additional tool to evaluate stress, given the existing capture and handling protocol for this species. Future studies should complete the evaluation of seasonal variation of cortisol and ACTH in unstressed individuals that was begun in the present study. Diurnal variation should also be examined, and was attempted in the present study, but could not be completed due to time constraints. Such studies are important because glucocorticoids and ACTH exhibit a circadian rhythm in most species, wherein concentrations are low while an animal is active, and rise (5-10x) during sleep so that concentrations are highest in the morning, upon waking [147]. Such time-dependent variation could affect cortisol and ACTH concentrations observed in the blood. 120

PAGE 121

LIST OF REFERENCES [1] Rathbun GB, Reid JP, Bonde RK, Powell JA. Reproduction in free-ranging Florida manatees. In: OShea TJ, Ackerman BB, and Percival HF, editors. Population biology of the Florida manatee, Information and Technology Report 1. Washington D.C.: U.S. Department of the Interior, National Biological Service; 1995. p. 135-56. [2] Daels PF, Albrecht BA, Mohammed HO. Equine chorionic gonadotropin regulates luteal steroidogenesis in pregnant mares. Biol Reprod 1998; 59: 1062-8. [3] Kirpatrick J, Lasley BL, Shideler SE, Roser JF, Turner Jr JW. Noninstrumented immunoassay field tests for pregnancy detection in free-roaming feral horses. J Wildl Manage 1993; 57: 168-73. [4] McNeilly AS, Martin RD, Hodges JK, Smuts GL. Blood concentrations of gonadotrophins, prolactin and gonadal steroids in males and in non-pregnant and pregnant female African elephants (Loxodonta africana). J Reprod Fertil 1983; 67: 113-20. [5] Hess DL, Schmidt AM, Schmidt MJ. Reproductive cycle of the Asian elephant (Elephas maximus) in captivity. Biol Reprod 1983; 28: 767-73. [6] Niemuller C, Gray C, Cummings E, Liptrap RM. Plasma concentrations of immunoreactive relaxin activity and progesterone in the pregnant Asian elephant (Elephas maximus). Anim Reprod Sci 1998; 53: 119-31. [7] Kellogg ME, Burkett S, Dennis TR, Stone G, Gray BA, McGuire PM, Zori RT, Stanyon R. Chromosome painting in the manatee supports Afrotheria and Paenungulata. BMC Evol Biol 2007; 7: 6. [8] Pardini AT, O'Brien PCM, Fu B, Bonde RK, Elder FFB, Ferguson-Smith MA, Yang F, Robinson TJ. Chromosome painting among Proboscidea, Hyracoidea and Sirenia: support for Paenungulata (Afrotheria, Mammalia) but not Tethytheria. Proc Royal Soc B: Biolog Sci 2007; 274: 1333-40. [9] Boyd JS, Omran SN, Ayliffe TR. Evaluation of real time B-mode ultrasound scanning for detecting early pregnancy in cows. Vet Rec 1990; 127: 350-2. [10] Engeland IV, Ropstad E, Andresen Eik LO. Pregnancy diagnosis in dairy goats using progesterone assay kits and oestrous observation. Anim Repro Sci 197; 47: 237-43. [11] Karen A, Beckers J-F, Sulon J, de Sousa NM, Szabados K, Reczigel J, Szenci O. Early pregnancy diagnosis in sheep by progesterone and pregnancy-associated glycoprotein tests. Theriogenology 2003; 59: 1941-8. 121

PAGE 122

[12] Larkin IL. Reproductive endocrinology of the Florida manatee (Trichechus manatus latirostris): estrous cycles, seasonal patterns and behavior. Ph.D. Dissertation. Gainesville, Florida: University of Florida; 2000. p. 40, 44, 48, 55, 56, 78, 103. [13] Dionysius DA. Pregnancy diagnosis in dairy goats and cows using progesterone assay kits. Aust Vet J 1991; 68: 14-6. [14] Raggi LA, Ferrando G, Parraguez VH, MacNiven V, Urquieta B. Plasma progesterone in alpaca (Lama pacos) during pregnancy, parturition and early postpartum. Anim Repro Sci 1999; 54: 245-9. [15] Chadio S, Xylouri E, Kalogiannis D, Michalopoulou E, Evagelatos S, Menegatos I. Early pregnancy diagnosis in swine by direct radioimmunoassay for progesterone in blood spotted on filter paper. Anim Repro Sci 2002; 69: 65-72. [16] Meyer HHD, Jewgenow K, Hodges JK. Binding activity of 5-reduced gestagens to the progestin receptor from African elephant (Loxodonta africana). Gen Comp Endocrinol 1997; 105: 164-67. [17] Hodges JK, Heistermann M, Beard A, van Aarde RJ. Concentrations of progesterone and the 5-reduced progestins, 5-pregnane-3,20-dione and 3-hydroxy-5-pregnan-20-one, in luteal tissue and circulating blood and their relationship to luteal function in the African elephant, Loxodonta africana. Biol Reprod 1997; 56: 640-6. [18] Hodges JK. Endocrinology of the ovarian cycle and pregnancy in the Asian (Elephas maximus) and African (Loxodonta africana) elephant. Anim Reprod Sci 1998; 53: 3-18. [19] Holtan DW, Houghton E, Silver M, Fowden AL, Ousey J, Rossdale PD. Plasma progestagens in the mare, fetus and newborn foal. J Reprod Fertil Suppl 1991; 44: 51728. [20] Smith MF, McIntush EW, Smith GW. Mechanisms associated with corpus luteum development. J Anim Sci 1994; 72: 1857-72. [21] Niswender GD, Juengel JL, Silva PJ, Rollyson MK, McIntush EW. Mechanisms controlling the function and life span of the corpus luteum. Physiol Rev 2000; 80: 1-29. [22] Anderson ML, Long JA. 1978. Localization of relaxin in the pregnant rat. Bioassay of tissue extracts and cell fractionation studies. Biol Reprod 1978; 18: 110-7. [23] Fields MJ, Fields PA. Morphological characteristics of the bovine corpus luteum during the estrous cycle and pregnancy. Theriogenology 1996; 45: 1295-325. 122

PAGE 123

[24] Marmontel M. Age and reproduction in female Florida manatees. In: OShea TJ, Ackerman BB, and Percival HF, editors. Population biology of the Florida manatee, Information and Technology Report 1. Washington D.C.: U.S. Department of the Interior, National Biological Service; 1995. p. 98-119. [25] Allen WR. Luteal deficiency and embryo mortality in the mare. Reprod Dom Anim 2001; 36: 121-31. [26] OShea TJ, Rathbun GB, Asper ED, Searles SW. Tolerance of West Indian manatees to capture and handling. Biol Conserv 1985; 33: 335-50. [27] Reeder DM, Kramer KM. Stress in free-ranging mammals: integrating physiology, ecology, and natural history. J Mammal 2005; 86: 225-35. [28] Millspaugh JJ, Washburn BE. Use of fecal glucocorticoid metabolite measures in conservation biology research: considerations for application and interpretation. Gen Comp Endocrinol 2004; 138: 189-99. [29] Cooke PS, Hoslberger DR, Witorsch RJ, Sylvester PW, Meredith JM, Treinen KA, Chapin RE. Thyroid hormone, glucocorticoids, and prolactin at the nexus of physiology, reproduction, and toxicology. Toxicol Appl Pharmacol 2004; 194: 309-35. [30] Mstl E, Palme R. Hormones as indicators of stress. Domest Anim Endocrinol 2002; 23: 67-74. [31] Gulland FMD, Haulena M, Lowenstine LJ, Munro C, Graham PA, Bauman J, Harvey J. Adrenal function in wild and rehabilitated Pacific harbor seals (Phoca vitulina richardii) and in seals with phocine herpesvirus-associated adrenal necrosis. Mar Mamm Sci 1999; 15: 810-27. [32] Rotllant J, Tort L. Cortisol and glucose responses after acute stress by net handling in the sparid red porgy previously subjected to crowding stress. J Fish Biol 1997; 51: 21-8. [33] Ruane NM, Huisman EA, Komen J. Plasma cortisol and metabolite level profiles in two isogenic strains of common carp during confinement. J Fish Biol 2001; 59: 1-12. [34] Suzuki J, Ohkura S, Terao K. Baseline and stress levels of cortisol in conscious and unrestrained Japanese macaques (Macaca fuscata). J Med Primatol 2002; 31: 340-4. [35] Munck A, Guyer PM, Holbrook NJ. Physiological functions of glucocorticoids in stress and their relation to pharmacological actions. Endocr Rev 1984; 5: 25-44. [36] Breazile JE. Physiologic basis and consequences of distress in animals. J Am Vet Med Assoc 1987; 191: 1212-5. 123

PAGE 124

[37] Turner JW Jr., Tolson P, Hamad N. Remote assessment of stress in white rhinoceros (Ceratotherium simum) and black rhinoceros (Diceros bicornis) by measurement of adrenal steroids in feces. J Zoo Wildl Med 2002; 33: 214-21. [38] Kranendonk G, Hoopster H, Fillerup M, Ekkel ED, Mulder EJH, Wiegant VM, Taverne MAM. Lower birth weight and attenuated adrenocortical response to ACTH in offspring from sows that orally received cortisol during gestation. Domest Anim Endocrinol 2006; 30: 218-38. [39] Turner AI, Hemsworth PH, Tilbrook AJ. Susceptibility of reproduction in female pigs to impairment by stress or elevation of cortisol. Domest Anim Endocrinol 2005; 29: 398410. [40] Francis-Floyd R, White JR, Chen CL, Cardeilhac PT, Cichra CE. Serum progesterone and estradiol concentrations in captive manatees Trichechus manatus. J Aquat Anim Health 1991; 3: 70-73. [41] Palme R. Measuring fecal steroids: guidelines for practical application. Ann NY Acad Sci 2005; 1046: 75-80. [42] Larkin IL, Gross TS, Reep RL. Use of faecal testosterone concentrations to monitor male Florida manatee (Trichechus manatus latirostris) reproductive status. Aq Mamm 2005; 31: 52-61. [43] Brown JL, Lehnhardt J. Serum and urinary hormones during pregnancy and the periand postpartum period in an Asian elephant (Elephas maximus). Zoo Biol 1995; 14: 555-64. [44] Brown JL. Reproductive endocrine monitoring of elephants: an essential tool for assisting captive management. Zoo Biol 2000; 19: 347-67. [45] Weigle BL, Wright IE, Ross M, Flamm R. Movements of radio-tagged manatees in Tampa Bay and along Floridas west coast, 1991-1996. Technical Report. St. Petersburg, Florida: Florida Marine Research Institute; 2001. p.3. [46] Deutsch CJ, Edwards HH, Barlas ME. Florida manatee foraging behavior around a winter warm-water refuge. Final Report to the Wildlife Foundation of Florida. St. Petersburg, Florida: FWC Fish and Wildlife Research Institute; 2006. 42pp. [47] Slone DH, Stith BM, Reid JP. Manatee use within Everglades National Park: data summary and management implications. Abstract only. Greater Everglades Ecosystem Restoration Conference. Lake Buena Vista, Florida. 5-9 June 2006. p. 210. [48] Morales-Vela B, Olivera-Gmez D, Reynolds III JE, Rathbun GB. Distribution and habitat use by manatees (Trichechus manatus manatus) in Belize and Chetumal Bay, Mexico. Biol Conserv 2000; 95: 67-75. 124

PAGE 125

[49] Reid JP. Cooperative manatee research in Puerto Rico. Endangered Species Bulletin 2006; 31: 18-9. [50] Marmontel M. Age determination and population biology of the Florida manatee, Trichechus manatus latirostris. Ph.D. Dissertation. Gainesville, Florida: University of Florida; 1993. p. 144, 172, 185. [51] Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin Chem 1993; 39: 561-77. [52] Munsterjhelm C, Valros A, Heinonen M, Hlli O, Peltoniemi OAT. Welfare index and reproductive performance in the sow. Reprod Dom Anim 2006; 41: 494-500. [53] Harr K, Harvey J, Bonde R, Murphy D, Lowe M, Menchaca M, Haubold E, Francis-Floyd R. Comparison of methods used to diagnose generalized inflammatory disease in manatees (Trichechus manatus latirostris). J Zoo Wildl Med 2006; 37: 151-9. [54] Chrousos GP, Torpy DJ, Gold PW. Interactions between the hypothalamic-pituitaryadrenal axis and the female reproductive system: clinical implications. Ann Intern Med 1998; 129: 229-40. [55] Agnew D, Munson L, Ramsay E. Cystic endometrial hyperplasia in elephants. Vet Path 2004; 41: 179-83. [56] Cornbleet PJ, Gochman N. Incorrect least-squares regression coefficients in methodcomparison analysis. Clin Chem 1979; 25: 432-8. [57] Chinchilli VM, Miller WG. Evaluating test methods by estimating total error. Clin Chem 1994; 40: 464-71. [58] Royal College of Pathologists of Australasia. Appendix A: Programs, analytes and allowable limits of performance. Analytical Quality Requirements in Chemical Pathology Quality Assurance Programs Information Handbook. Adelaide, Australia: Australasian Association of Clinical Biochemists; 2001. p. 27. [59] Westgard JO, Carey RN, Wold S. Criteria for judging precision and accuracy in method development and evaluation. Clin Chem 1974; 20: 825-33. [60] Stckl D, Dewitte K, Thienpont LM. Validity of linear regression in method comparison studies: is it limited by the statistical model or the quality of the analytical input data? Clin Chem 1998; 44: 2340-6. [61] Comstock, GW, Burke AE, Norkus EP, Gordon GB, Hoffman SC, Helzlsouer KJ. Effects of repeated freeze-thaw cycles on concentrations of cholesterol, micronutrients, and hormones in human plasma and serum. Clin Chem 2001; 47: 139-42. 125

PAGE 126

[62] de Villiers DJ, Skinner JD, Hall-Martin AJ. Circulating progesterone concentrations and ovarian functional anatomy in the African elephant (Loxodonta africana). J Reprod Fertil 1989; 86: 195-201. [63] LeBlanc MM. Equine perinatology: what we know and what we need to know. Anim Reprod Sci 1996; 42: 189-96. [64] Joachim RA, Hildebrandt M, Oder J, Klapp BF, Arck PC. Murine stress-triggered abortion is mediated by increase of CD8+ TNF-+ decidual cells via substance P. Am J Reprod Immunol 2001; 45: 303-9. [65] OShea TJ, Hartley WC. Reproduction and early-age survival of manatees at Blue Spring, upper St. Johns River, Florida. In: OShea TJ, Ackerman BB, and Percival HF, editors. Population biology of the Florida manatee, Information and Technology Report 1. Washington D.C.: U.S. Department of the Interior, National Biological Service; 1995. p. 157-70. [66] Ackerman BB, Wright SD, Bonde RK, Odell DK, Banowetz DJ. Trends and patterns in mortality of manatees in Florida, 1974-1991. In: OShea TJ, Ackerman BB, and Percival HF, editors. Population biology of the Florida manatee, Information and Technology Report 1. Washington D.C.: U.S. Department of the Interior, National Biological Service; 1995. p. 223-58. [67] Hernandez P, Reynolds III JE, Marsh H, Marmontel M. Age and seasonality in spermatogenesis of Florida manatees. In: OShea TJ, Ackerman BB, and Percival HF, editors. Population biology of the Florida manatee, Information and Technology Report 1. Washington D.C.: U.S. Department of the Interior, National Biological Service; 1995. p. 84-95. [68] Concannon PW, Powers ME, Holder W, Hansel W. Pregnancy and parturition in the bitch. Biol Reprod 1977; 16: 517-26. [69] Dehnhard M, Hildebrandt T, Rohleder M, Strauss G, Meyer HH, Goritz F. Application of an enzyme-immunoassay (EIA) for rapid screening of 5 alpha-pregnane-3,20-dione (DHP) in blood plasma of the Asian elephant, Elephas maximus. Berl Munch Tierarztl Wochenschr 2001; 114: 161-5. [70] Lu JKH, Judd HL. Silastic implants of progesterone produce high circulating levels of both progesterone and 20-hydroxyprogesterone in ovarectomized, adrenalectomized rats. Biol Reprod 1982; 26: 385-90. [71] Tsumagari S, Kamata J, Takagi K, Tanemura K, Yosai A, Takeishi M. 3 betahydroxysteroid dehydrogenase activity and gestagen concentrations in bovine cotyledons and caruncles during gestation and parturition. J Reprod Fertil 1994; 102: 35-9. 126

PAGE 127

[72] Gambelunghe C, Rossi R, Ferranti C, Rossi R, Bacci M. Hair analysis by GC/MS/MS to verify abuse of drugs. J Appl Toxicol 2005; 25: 205-11. [73] Sparkman OD. Mass Spectrometry PittCon 2001. J Am Soc Mass Spectrom 2001; 12: 607-16. [74] Kirkman S, Wallace ED, van Aarde RJ, Potgieter HC. Steroidogenic correlates of pregnancy in the rock hyrax (Procavia capensis). Life Sciences 2001; 68: 2061-72. [75] Rommel SA, Pabst DA, McLellan WA. Functional morphology of venous structures associated with the male and female reproductive systems in Florida manatees (Trichechus manatus latirostris). Anat Rec 2001; 264: 339-47. [76] Cicinelli E, de Ziegler D. Transvaginal progesterone: evidence for a new functional portal system flowing from the vagina to the uterus. Hum Reprod Update 1999; 5: 365-72. [77] Einer-Jensen N, Hunter RHF. Counter-current transfer in reproductive biology. Reproduction 2005; 129: 9-18. [78] Jewgenow K, Meyer HHD. Comparative binding affinity study of progestins to the cytosol progestin receptor of endometrium in different mammals. Gen Comp Endo 1998; 110: 118-24. [79] Ousey JC, Forhead AJ, Rossdale PD, Grainger L, Houghton E, Fowden AL. Ontogeny of uteroplacental progestagen production in pregnant mares during the second half of gestation. Biol Reprod 2003; 69: 540-8. [80] Wilcox RB, Wiest WG. Comparative effectiveness of progesterone and 4-pregnen-20-ol3-one in the development of deciduomata. Endocrinol 1960; 67: 281-2. [81] Illingworth DV, Heap RB, Perry JS. Changes in the metabolic clearance rate of progesterone in the guinea-pig. J Endocrinol 1970; 48: 409-17. [82] Illingworth DV, Heap RB. A decrease in the metabolic clearance rate of progesterone in the coypu during pregnancy. J Reprod Fertil 1971; 27: 492-4. [83] Marsh H. Mass stranding of dugongs by a tropical cyclone. Mar Mamm Sci 1989; 5: 7584. [84] Klima F, Rohleder M, Dehnhard M, Meyer HHD. Identification of progesterone and 5dihydroprogesterone (5-DHP) in zebras by two-dimensional high-performance liquid chromatography and their dependence on the state of reproduction. Zoo Biol 1999; 18: 325-33. 127

PAGE 128

[85] Fraser HM, Lunn SF, Harrison DJ, Kerr JB. Luteal regression in the primate: different forms of cell death during natural and gonadotropin-releasing hormone antagonist or prostaglandin analogue-induced luteolysis. Biol Reprod 1999; 61: 1468-79. [86] Townson DH. Immune cell-endothelial cell interactions in the bovine corpus luteum. Integr Comp Biol 2006; 46: 1055-9. [87] Fitz TA, Mayan MH, Sawyer HR, Niswender GD. Characterization of two steroidogenic cell types in the ovine corpus luteum. Biol Reprod 1982; 27: 703-11. [88] Jablonka-Shariff A, Grazul-Bilska AT, Redmer DA, Reynolds LP. Growth and cellular proliferation of the ovine corpora lutea throughout the estrous cycle. Endocrinol 1993; 133: 1871-9. [89] Friedman A, Weiss S, Levy N, Meidan R. Role of tumor necrosis factor and its type I receptor in luteal regression: induction of programmed cell death in bovine corpus luteum-derived endothelial cells. Biol Reprod 2000; 63: 1905-12. [90] Niswender GD, Nett TM. The corpus luteum and its control. In: Knobil E, Neill J, editors. The Physiology of Reproduction. New York, New York: Raven Press; 1988. p. 489-525. [91] Mossman HW, Duke KL. Comparative morphology of the mammalian ovary. Madison, Wisconsin: University of Wisconsin Press; 1973. p. 216, 269-71. [92] Peluso JJ, Pappalardo A, Trolice MP. N-cadherin-mediated cell contact inhibits granulosa cell apoptosis in a progesterone-independent manner. Endocrinol 1996; 137: 1196-203. [93] Roughton SA, Lareu RR, Bittles AH, Dharmarajan AM. Fas and fas ligand messenger ribonucleic acid and protein expression in the rat corpus luteum during apoptosismediated luteolysis. Biol Reprod 1999; 60: 797-804. [94] Paavola LG. The corpus luteum of the guinea pig III. Cytochemical studies on the Golgi complex and GERL during normal postpartum regression of luteal cells, emphasizing the origin of lysosomes and autophagic vacuoles. J Cell Biol 1978; 79: 59-73. [95] Levine H, Wight T, Squires E. Ultrastructure of the corpus luteum of the cycling mare. Biol Reprod 1979; 20: 492-504. [96] Bulling, A, Berg FD, Berg U, Duffy DM, Stouffer RL, Ojeda SR, Gratzl M, Mayerhofer A. Identification of an ovarian voltage-activated Na+ -channel type: hints to involvement in luteolysis. Mol Endocrinol 2000; 14: 1064-74. [97] Gemmell RT, Stacy BD, Thorburn GD. Ultrastructural study of secretory granules in the corpus luteum of the sheep during the estrous cycle. Biol Reprod 1976; 11: 447-62. 128

PAGE 129

[98] OShea JD, Cran DG, Hay MF. The small luteal cell of the sheep. J Anat 1979; 128: 23951. [99] Fields MJ, Dubois W, Fields PA. Dynamic features of luteal secretory granules: ultrastructural changes during the course of pregnancy in the cow. Endocrinol 1985; 117: 1675-82. [100] Fields MJ, Barros CM, Watkins WB, Fields PA. Characterization of large luteal cells and their secretory granules throughout the estrous cycle of the cow. Biol Reprod 1992; 46: 535-45. [101] Deane, HW, Hay MF, Moor RM, Rowson LEA, Short RV. The corpus luteum of the sheep: relationships between morphology and function during the oestrous cycle. Acta Endocrinol (Kbh) 1966; 51: 245-63. [102] Nett TM, McClellan MC, Niswender GD. Effects of prostaglandins on the ovine corpus luteum: blood flow, secretion of progesterone and morphology. Biol Reprod 1976; 15: 66-78. [103] Guo K, Wolf V, Dharmarajan AM, Feng Z, Bielke W, Saurer S, Friis R. Apoptosisassociated gene expression in the corpus luteum of the rat. Biol Reprod 1998; 58: 739-46. [104] Marmontel M. The reproductive anatomy of the female manatee Trichechus manatus latirostris (Linnaeus 1758) based on gross and histologic observations. M.Sc. Thesis. Coral Gables, Florida: University of Miami; 1988. p. 30, 34, 35. 38, 58, 59, 75, 76. [105] Smith NS, Buss IO. Formation, function, and persistence of the corpora lutea of the African elephant (Loxodonta africana). J Mamm 1975; 56: 30-43. [106] Hanks J, Short RV. The formation and function of the corpus luteum in the African elephant, Loxodonta africana. J Reprod Fert 1972; 29: 79-89. [107] Buss IO, Smith NS. Observations on reproduction and breeding behavior of the African elephant. J Wildl Manage 1966; 30: 375-88. [108] Perry JS. The reproduction of the African elephant, Loxodonta africana. Phil Trans Royal Soc London. Series B, Biolog Sci 1953; 237: 93-149. [109] Garcia A, Salaheddine M. Ultrasonic morphology of the corpora lutea and central luteal cavities during selection of recipients for embryo transfer. Reprod Dom Anim 2000; 35: 113-8. [110] Kimura J, Tsukise A, Yokota H, Nambo Y, Higuchi T. The application of threedimensional internal structure microscopy in the observation of mare ovary. Anat Histol Embryol 2001; 30: 309-12. 129

PAGE 130

[111] Marsh H, Heinsohn GE, Marsh LM. Breeding cycle, life history and population dynamics of the dugong, Dugong dugon (Sirenia: Dugongidae). Aust J Zool 1984; 32: 767-802. [112] Laws RM. Aspects of reproduction in the African elephant, Loxodonta africana. J Reprod Fertil, Suppl 1969; 6: 193-217. [113] Cole HH, Howell CE, Hart GH. The changes occurring in the ovary of the mare during pregnancy. Anat Rec 1931; 49: 199-209. [114] Amoroso EC, Hancock JL, Rowlands IW. Ovarian activity in the pregnant mare. Nature 1948; 161: 355-6. [115] Amoroso EC, Rowlands IW. Hormonal effects in the pregnant mare and foetal foal. J Endocrinol 1951; 7: 1-1ii. [116] ODonoghue PN. Reproduction in the female hyrax (Dendrohyrax arborea ruwenzorii). Proc Zool Soc Lond 1962; 141: 207-37 + 6 plates. [117] OShea JD, Rodgers RJ, DOcchio MJ. Cellular composition of the cyclic corpus luteum of the cow. J Reprod Fertil 1989; 85: 483-7. [118] Farin CE, Moeller CL, Sawyer HR, Gamboni F, Niswender GD. Morphometric analysis of cell types in the ovine corpus luteum throughout the estrous cycle. Biol Reprod 1986; 35: 1299-308. [119] Braden TD, Gamboni F, Niswender GD. Effects of prostaglandin F 2 -induced luteolysis on the populations of cells in the ovine corpus luteum. Biol Reprod 1988; 39: 245-53. [120] Ogle TF, Braach HH, Buss IO. Fine structure and progesterone concentration in the corpus luteum of the African elephant. Anat Rec 1973; 175: 707-24. [121] Niswender GD, Schwall RH, Fitz TA, Farin CE, Sawyer HR. Regulation of luteal function in domestic ruminants; new concepts. Recent Prog Horm Res 1985; 41: 101-51. [122] Rueda BR, Hendry IR, Hendry III WJ, Stormshak F, Slayden OD, Davis JS. Decreased progesterone levels and progesterone receptor antagonists promote apoptotic cell death in bovine luteal cells. Biol Reprod 2000; 62: 269-76. [123] Kayanja FIB, Sale JB. The ovary of rock hyrax of the genus Procavia. J Reprod Fert 1973; 33: 223-30. [124] Grizzi F, Chiriva-Internati M. Human binucleate hepatocytes: are they a defence during chronic liver diseases? Med Hypoth 2007; 69: 258-61. [125] Reynolds LP, Killilea SD, Redmer DA. Angiogenesis in the female reproductive system. The FASEB J 1992; 6: 886-92. 130

PAGE 131

[126] Fraser HM, Wulff C. Angiogenesis in the primate ovary. Reprod Fertil Dev 2001; 13: 557-66. [127] Kendall JZ, Plopper CG, Bryant-Greenwood GD. Ultrastructural immunoperoxidase demonstration of relaxin in corpora lutea from a pregnant sow. Biol Reprod 1978; 18: 948. [128] Tripp KM. Functional morphology of the gross anatomy of the ovaries of the Florida manatee (Trichechus manatus latirostris). B.S. Honors Thesis. St. Petersburg, Florida: Eckerd College; 2003. p. 21-23, 41-43, 45, 46. [129] Green MP, Mann GE, Hunter MG. Luteal characteristics and progesterone production on day 5 of the bovine oestrous cycle. Reprod Dom Anim 2007; 42: 643-7. [130] Eker K, Salmanolu MR. Ultrasonographic monitoring of follicular development, ovulation and corpora lutea formation in a bitch. Turk J Vet Anim Sci 2006; 30: 589-92. [131] Youngquist RS. Current Therapy in Large Animal Theriogenology. Philadelphia, Pennsylvania: W.B. Saunders; 1997. p. 301, 506, 519-520. [132] Ousey JC. Peripartal endocrinology in the mare and foetus. Reprod Domest Anim 2004; 39: 222-31. [133] Senger PL. Pathways to Pregnancy and Parturition, 2nd ed., Pullman, Washington: Current Conceptions, Inc.; 2005. p. 317. [134] Sugawara T, Hoshi N, Sakuragi N. Immunohistochemical study of steroidogenic acute regulatory (StAR) protein in ovarian tumors. Internet J Gynecol Obstet 2004; 3. [135] Stocco DM. StAR protein and regulation of steroid hormone biosynthesis. Ann Rev Phys 2001; 63: 193-213. [136] Christenson LK, Devoto L. Cholesterol transport and steroidogenesis by the corpus luteum. Reprod Biol Endocrinol 2003; 1: 90. [137] Pollack SE, Furth EE, Kallen BB, Arakane F, Kiriakidou M, Kozarsky KF, Strauss III JF. Localization of the steroidogenic acute regulatory protein in human tissues. J Clin Endocrinol Metab 1997; 82: 4243-51. [138] Devoto L, Kohen P, Gonzalez RR, Castro O, Retamales I, Vega M, Carvallo P, Christenson LK, Strauss JF. Expression of steroidogenic acute regulatory protein in the human corpus luteum throughout the luteal phase. J Clin Endocrinol Metab 2001; 86: 5633-9. 131

PAGE 132

[139] Dudek D. Can macroscopic identification of large ovarian follicles be useful in eliminating follicles with impaired steroid function? Cytotechnology 1994; 14: 147-53. [140] Marsh H, Heinsohn GE, Channels PW. Changes in the ovaries and uterus of the dugong, Dugong dugon (Sirenia: Dugongidae), with age and reproductive activity. Aust J Zool 1984; 32: 743-66. [141] Czekala NM, MacDonald EA, Steinman K, Walker S, Garrigues II NW, Olson D, Brown JL. Estrogen and LH dynamics during the follicular phase of the estrous cycle in the Asian elephant. Zoo Biol 1994; 22: 443-54. [142] Reid JP, Bonde RK, OShea TJ. Reproduction and mortality of radio-tagged and recognizable manatees on the Atlantic coast of Florida. In: OShea TJ, Ackerman BB, and Percival HF, editors. Population biology of the Florida manatee, Information and Technology Report 1. Washington D.C.: U.S. Department of the Interior, National Biological Service; 1995. p. 171-91. [143] Hartman D. Ecology and Behavior of the Manatee (Trichechus manatus) in Florida. Special Publication No. 5. The American Society of Mammalogists, Ithaca, New York; 1979. p. 95. [144] Ackerman BB. Aerial surveys of manatees: a summary and progress report. In: OShea TJ, Ackerman BB, and Percival HF, editors. Population biology of the Florida manatee, Information and Technology Report 1. Washington D.C.: U.S. Department of the Interior, National Biological Service; 1995. p. 13-33. [145] Anderson PK. Habitat, niche, and evolution of Sirenian mating systems. J Mamm Evol 2002; 9: 55-98. [146] Reynolds JE, Rommel SA, Pitchford ME. The likelihood of sperm competition in manatees-explaining an apparent paradox. Mar Mamm Sci 2004; 20: 464-76. [147] Touma C, Palme R. Measuring fecal glucocorticoid metabolites in mammals and birds: the importance of validation. Ann NY Acad Sci 2005; 1046: 54-74. [148] Ortiz RM, MacKenzie DS, Worthy GAJ. Osmoregulation in wild and captive West Indian manatees (Trichechus manatus). Physiol Zool 1998; 71: 449-57. [149] Manire CA, Walsh CJ, Rhinehart HL, Colbert DE, Noyes DR, Luer CA. Alterations in blood and urine parameters in two Florida manatees (Trichechus manatus latirostris) from simulated conditions of release following rehabilitation. Zoo Biol 2003; 22: 103-20. [150] Reisch N, Reincke M, Bidlingmaier M. Preanalytical stability of adrenocorticotropic hormone depends on time to centrifugation rather than temperature. Clin Chem 2007; 53: 358-9. 132

PAGE 133

[151] Wong A. Monitoring oral temperature, heart rate, and respiration rate of field-captured Florida and Antillean manatees (Trichechus manatus latirostris and T. m. manatus). Masters Thesis. Gainesville, Florida: University of Florida; 2007. p. 40, 43, 44. [152] Ortiz RM, Wade CE, Ortiz CL. Effects of prolonged fasting on plasma cortisol and TH in postweaned northern elephant seal pups. Am J Physiol Regul Integr Comp Physiol 2001; 280: R790-5. [153] St. Aubin DJ, Deguise S, Richard PR, Smith TG, Geraci JR. Hematology and plasma chemistry as indicators of health and ecological status in beluga whales, Delphinapterus leucas. Arctic 2001; 54: 317-31. [154] Sanvito S, Galimberti F, Delahunty KM, McKay DW. Blood spots in Pinnipedia hormone studies: measure of cortisol levels in southern elephant seals (Mirounga leonina). Aq Mamm 2004; 30: 251-6. [155] Good T, Khan MZ, Lynch JW. Biochemical and physiological validation of a corticosteroid radioimmunoassay for plasma and fecal samples in oldfield mice (Peromyscus polionotus). Physiol Behav 2003; 80: 405-11. [156] Bossart GD, Meisner RA, Rommel SA, Ghim S, Jenson AB. Pathological features of the Florida manatee cold stress syndrome. Aq Mamm 2002; 29: 9-17. [157] Brown JL, Wemmer CM, Lehnhardt J. Urinary cortisol analysis for monitoring adrenal activity in elephants. Zoo Biol 2005; 14: 533-42. [158] Yon L, Kanchanapangka S, Chaiyabutr N, Stanczyk FZ, Meepan S, Lasley B. ACTH stimulation in four Asian bull elephants (Elephas maximus): an investigation of androgen sources in bull elephants. Gen Comp Endocrinol 2007; 151: 246-51. [159] Bechert US, Swanson L, Wasser SK, Hess DL, Stormshak F. Serum prolactin concentrations in the captive female African elephant: potential effects of season and steroid hormone interactions. Gen Comp Endocrinol 1999; 114: 269-78. [160] Harvey JW, Harr KE, Murphy D, Walsh MT, Chittick EJ, Bonde RK, Pate MG, Deutsch CJ, Edwards HH, Haubold EM. Clinical biochemistry in healthy manatees (Trichechus manatus latirostris). J Zoo Wildl Med 2007; 38: 269-79. [161] Becker J, Craig EA. Heat shock proteins as molecular chaperones. Eur J Biochem 1994; 219: 11-23. [162] Feder ME. Organismal ecological and evolutionary aspects of heat-shock proteins and the stress response. Amer Zool 1999; 39: 857-64. [163] Iwama GK, Vijayan MM, Forsyth RB, Ackerman PA. Heat shock proteins and physiological stress in fish. Amer Zool 1999; 39: 901-9. 133

PAGE 134

[164] Zgel U, Kaufmann SHE. Role of heat shock proteins in protection from and pathogenesis of infectious diseases. Clin Microbiol Rev 1999; 12: 19-39. [165] Bechert US, Southern S. Monitoring environmental stress in African elephants through molecular analysis of stress-activated proteins. In: Baer CK, editor. Proceedings of the American Association of Zoological Veterinarians. Milwaukee, Wisconsin. 5-10 October 2002. p. 249-53. 134

PAGE 135

BIOGRAPHICAL SKETCH Katie Tripp was born and raised in Harveys Lake, Pennsylvania. In August 1999, she moved to St. Petersburg, Florida to attend Eckerd College, where she majored in marine science and double-minored in chemistry and environmental policy. Katie completed her B.S. with honors in May 2003. She began her Ph.D. studies with UFs Aquatic Animal Health Program in January 2004. 135