Gene regulation in teleosts by estradiol and estrogen mimics

MISSING IMAGE

Material Information

Title:
Gene regulation in teleosts by estradiol and estrogen mimics
Physical Description:
Unknown
Creator:
Bowman, Christopher James
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 26576874
oclc - 70833743
System ID:
AA00022167:00001

Full Text









GENE REGULATION IN TELEOSTS BY ESTRADIOL AND ESTROGEN MIMICS


By

CHRISTOPHER JAMES BOWMAN











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


2001














ACKNOWLEDGMENTS


I would like to thank the many people who contributed to my work over the past

four years. Their help and advice provided me with the necessary training to conduct

these studies, and the importance of patience and collaboration while performing sound

scientific research.

Nancy D. Denslow, Ph.D. showed tremendous compassion and incredible support

while I was in her laboratory. She introduced me to the field of endocrine disruption, and

together we learned more about environmental toxicology. She provided a place to work

with seemingly unlimited resources. She also offered an environment of cutting-edge

biotechnology with unlimited possibilities. Nancy freely shared ideas and advice, yet

allowed me to decide what projects to work on and how to complete them, this

encouraged and fostered my independence as a researcher. I feel like together we were

able to write fellowship applications that have financially supported me for the past four

years. With her support, I had the opportunity to share my work with many different

scientific audiences, teaching me the value of effective communication. Nancy embodies

many desirable qualities that are rare in a doctoral advisor and I feel blessed to be her

first doctoral student.

Collaboration with Leroy Folmar, Ph.D., Michael Hemmer, Ph.D., and their

colleagues at the U.S. E.P.A. (United States Environmental Protection Agency) was

educational and successful. The experiments and discussions we shared led to many








publications and an entire chapter of this dissertation. I thank them all for the

contributions they made to my doctoral research.

I would also like to thank my committee members: Kathleen Shiverick, Ph.D.,

Stephen Roberts, Ph.D., Evan Gallagher, Ph.D., and Rosalia Simmen, Ph.D. They

provided outstanding support and guidance throughout the research process and I thank

them for reviewing this dissertation.

Kevin Kroll, Marjorie Chow, Ming Chow, Scott McClung, Scott McMillan,

Alfred Chung, Li Zhang, Luan Le, and Ben O'Neal of the Molecular Biomarkers and

Protein Chemistry Core labs are thanked for all the technical guidance and support during

my time in the Denslow laboratories. I especially appreciate Kevin and Marjorie's hard

work and patience. Without them the work would have taken much longer than it did.

I would like to thank Sandy Gibbons, Teresa Stevens, and Robert Coney in the

ICBR (Interdisciplinary Center for Biotechnology Research) for keeping up with my

never-ending requests. Ron Ferguson, Tammy Flagg, Steve Lee, Linda Green, Karen

Kelley, Scott Whittaker, David Moraga, Sharon Norton, Keith Lowe, William Farmerie,

and Savita Shanker all provided valuable assistance through their technical expertise and

advice. I would especially like to thank Ron, Sharon, and Steve who first taught me the

basics of molecular biology. I appreciate all their help. Without them this research

would have taken much longer than it did.

The interdisciplinary nature of my dissertation required administrative support

from several departments. I thank Judy Adams and the staff in the Department of

Pharmacology for their unconditional care and understanding while explaining the

University system. I would also like to thank the administrative staff in the Center for








Environmental and Human Toxicology, ICBR, and the Department of Biochemistry and

Molecular Biology for all their assistance.

I extend much appreciation to Evan Gallagher and people from his lab (Karen

Pastos, Kristen Henson, and James Gardner) who provided significant technical and

scientific advice to this research. I would also like to thank Greg Stauffer and Kathy

Childress for their help in the Aquatic Toxicology lab. I will miss our joint lab meetings.

I thank Timothy Gross and his staff at the United States Geological Survey

(Shane Ruessler, Carla Wieser, and Jon Wiebe) for providing facilities, personnel, and

fish for some of the experiments. I also greatly appreciate the plasma hormone analysis

conducted by Tim's lab.

This disseration would not have been possible without the financial support of the

Superfund Basic Research Program (National Institutes of Environmental Health

Sciences) Center grant and Graduate Fellowship. In addition, the U.S. EPA-STAR

(Science to Achieve Results) Graduate Fellowship I received provided support for my

salary, tuition, and limited expenses from 1999 to 2001.

Finally, with special recognition and love, I thank my wife, Robin Bowman, for

her unconditional support and patience. Our many years spent in Gainesville will never

be forgotten. I also thank close friends and family for their kind words and

encouragement. I would not have been able to finish this degree without them.














TABLE OF CONTENTS

p~ge

A CKN OW LED GM ENTS ............................................................................................ ii

LIST OF TAB LES ........................................................................................................... viii

LIST OF FIGURES ........................................................................................................ ix

ABSTRA CT ...................................................................................................................... xii

CHAPTERS

1 INTRODUCTION AND LITERATURE REVIEW ................................................. 1

Background ..................................................................................................................... 1
The Endocrine Disruption Hypothesis ........................................................................ 1
Supporting Environm ental Evidence .......................................................................... 1
Supporting Laboratory Evidence .......................................................................... 3
Xenobiotics w ith Estrogenic Activity ........................................................................ 5
Pharm aceutical Estrogens ...................................................................................... 6
Organochlorine Com pounds .................................................................................... 6
Phenolics ..................................................................................................................... 7
N aturally Occurring Com pounds ........................................................................... 8
Other Com pounds .................................................................................................... 8
Fish M odels ..................................................................................................................... 9
Vitellogenin ................................................................................................................... 11
M odel of Horm onal Regulation ............................................................................. 12
Biom arker of Estrogenicity .................................................................................... 12
Estrogen Receptor M ediated Pathw ay ...................................................................... 13

2 COORDINATE IN VIVO GENE EXPRESSION IN LARGEMOUTH BASS
(MICROPTERUS SALMOIDES) AFTER ACUTE ESTRADIOL EXPOSURE .......... 20

Introduction ................................................................................................................... 20
Coordinate Gene Regulation .................................................................................. 21
Largem outh Bass ................................................................................................. 23
Experim ental Objectives and Hypothesis ............................................................. 24
M aterials and M ethods ............................................................................................. 24
Fish Collection and M aintenance ......................................................................... 24
Experim ental Exposures ...................................................................................... 25








RN A Isolation ........................................................................................................... 26
Estrogen Receptor and V itellogenin RT-PCR ...................................................... 27
Cloning and Sequencing of Isolated cDN A s ........................................................ 27
m RN A Quantification ........................................................................................... 28
Differential D isplay RT-PCR ............................................................................... 30
Protein Analysis ................................................................................................... 31
Results ........................................................................................................................... 33
Estrogen Receptor and Vitellogenin Sequencing and Characterization ............... 33
Regulation of mRNAs in Adult Largemouth Bass 48 Hours Post Injection ...... 34
Plasm a V itellogenin D ose Response ................................................................... 36
mRNA Characterization After Acute Exposure in Juvenile Largemouth Bass ........ 37
Discussion ..................................................................................................................... 38
Largemouth Bass Estrogen Receptor and Vitellogenin ........................................ 39
Exposure of Adult Largemouth Bass to Estradiol ............................................... 42
Plasma Steroid and Vitellogenin Dose Response to Estrogens ............................ 44
Exposure of Juvenile Largemouth Bass to Estradiol .......................................... 48
Differential m RN A Regulation by Estradiol ........................................................ 53

3 DEVELOPMENT OF AN iN VITRO MODEL: LARGEMOUTH BASS PRIMARY
HEPA TOCYTES ...................................................................................................... 77

Introduction ................................................................................................................... 77
Rationale ................................................................................................................... 78
Fish Prim ary H epatocytes .................................................................................... 79
Objective ................................................................................................................... 81
M aterials and M ethods ............................................................................................. 82
Fish Collection and M aintenance ......................................................................... 82
Reagents and Equipm ent ...................................................................................... 82
Buffers and M edia ................................................................................................ 84
H epatocyte Isolation ................................................................................................. 85
The Fish .................................................................................................................... 86
The Perfusion ........................................................................................................ 86
The Isolation of H epatocytes ............................................................................... 88
H epatocyte Culture ................................................................................................... 89
Cell V iability ......................................................................................................... 89
Culture M orphology ............................................................................................. 91
Experim ental Results ............................................................................................ 91
Discussion ..................................................................................................................... 93

4 VITELLOGENIN INDUCTION IN SHEEPSHEAD MINNOW AS AN IN VIVO
M OD EL FOR ESTROGENICITY ............................................................................. 106

Introduction ................................................................................................................. 106
Vitellogenin mRNA and Protein as Biomarkers of Exposure ................................ 107
Environm ental Estrogens ........................................................................................ 108
Experim ental Objective .......................................................................................... 110
M aterials and M ethods ................................................................................................ 111








Fish Collection and M aintenance ............................................................................ 111
Experimental Exposures ......................................................................................... 112
RNA Isolation, Identification and Verification of Vitellogenin Sequences ........... 114
mRNA Quantification ............................................................................................. 115
Differential Display RT-PCR ................................................................................. 117
Protein Analysis ...................................................................................................... 118
Results ......................................................................................................................... 120
Vitellogenin I & II Cloning, Sequencing and Characterization .............................. 120
Time Course of Vitellogenin Induction Post Injection ........................................... 122
Vitellogenin Dose and Time Response with Constant Aqueous Exposure ............ 124
Gene regulation by estradiol, diethylstilbestrol, and ethinylestradiol ................. 125
Vitellogenin induction by nonylphenol, methoxychlor, and endosulfan ............ 126
Decreased vitellogenin levels following transfer to clean water ........................ 128
Discussion ................................................................................................................... 129
Acute Estradiol Exposures ...................................................................................... 132
Aqueous Dose Response Exposures ....................................................................... 135
Decreased Vitellogenin Levels After Exposure ...................................................... 140

5 CONCLUSIONS .......................................................................................................... 160

LIST OF REFERENCES ................................................................................................. 172

BIOGRAPHICAL SKETCH ........................................................................................... 196














LIST OF TABLES


Table Page

2-1. Plasma vitellogenin in largemouth bass 48 hours post estradiol injection ...... 56

3-1. Composition of buffered perfusion solutions used to isolate primary hepatocytes..97

3-2. Media and supplements used to culture primary hepatocytes ............................ 97

4-1. Nominal and actual measured water concentrations of chemicals evaluated in
sheepshead m innow ................................................................................................ 144

4-2. Vitellogenin mRNA and protein induction after seven days aqueous exposure .... 145














LIST OF FIGURES


Fige Page

1-1. Representative chemical structures of different families of suspected
xenoestrogens .................................................................................................... 17

1-2. Relationship between the liver and ovary during vitellogenesis in females .......... 18

1-3. Simplified scheme of the estrogen receptor mediated pathway ........................ 19

2-1. Acute mRNA response hypothesis .................................................................... 57

2-2. Largemouth bass (Micropterus salmoides) ...................................................... 57

2-3. Largemouth bass sampling by lab personnel ................................................... 58

2-4. Histology of largemouth bass gonads ............................................................... 59

2-5. Estrogen receptor and vitellogenin RT-PCR products ...................................... 60

2-6. Cloning strategy for estrogen receptor and vitellogenin PCR products ........... 61

2-7. Largemouth bass vitellogenin cRNA standard curve ....................................... 62

2-8. Multiple sequence alignment of cloned largemouth bass estrogen receptor
fragm ent ................................................................................................................. 63

2-9. Multiple sequence alignment of cloned largemouth bass vitellogenin fragment ..64

2-10. Largemouth bass liver mRNA characterization ............................................... 65

2-11. Plasma steroid measurements out to 48 hours post injection .......................... 66

2-12. Estradiol induced mRNAs over 48 hours ........................................................ 67

2-13. Plasma vitellogenin induction over 48 hours .................................................... 68

2-14. Largemouth bass differential display primer pair G-23 .................................... 69

2-15. Largemouth bass differential display using primer pairs G-10 and C-1 ........... 70








2-16. Verification of largemouth bass ERp72 ........................................................... 71

2-17. Plasma steroids 48 hours following acute exposure ........................................ 72

2-18. Dose response induction of plasma vitellogenin in largemouth bass ............... 73

2-19. Time course of estradiol induced mRNAs over 21 days ................................. 74

2-20. Time course of plasma vitellogenin induction ................................................. 75

2-21. Largemouth bass estrogen receptor and vitellogenin mRNA comparison ..... 76

3-1. Liver perfusion setup ........................................................................................ 98

3-2. Hepatocyte isolation cytology ........................................................................... 99

3-3. Electron microscopy of isolated hepatocytes ....................................................... 100

3-4. Trypan blue of isolated hepatocytes .................................................................... 101

3-5. Hepatocyte viability over time ............................................................................. 102

3-6. Osmolality of hepatocyte reagents ....................................................................... 103

3-7. Light microscopy of cultured hepatocytes ........................................................... 104

3-8. Primary hepatocyte gene expression .................................................................... 105

4-1. Sheepshead minnow (Cyprinidon variegatus) ..................................................... 145

4-2. RT-PCR on estrogen treated sheepshead minnow RNA ..................................... 146

4-3. Multiple sequence alignment of sheepshead minnow VIT 1 and VIT 2 ............. 146

4-4. Northern blot analysis of the two sheepshead minnow vitellogenin mRNAs ..... 147

4-5. Time course of vitellogenin mRNA induction in sheepshead minnow following
acute estradiol exposure ....................................................................................... 148

4-6. Time course of sheepshead minnow plasma vitellogenin induction following
acute estradiol exposure ....................................................................................... 149

4-7. Induction of sheepshead minnow vitellogenin mRNA and protein after seven days
constant exposure ................................................................................................. 150

4-8. Induction of sheepshead minnow vitellogenin mRNA following constant aqueous
exposure ............................................................................................................... 151








4-9. Induction of sheepshead minnow plasma vitellogenin following constant aqueous
exposure ............................................................................................................... 152

4-10. Sheepshead minnow differential display primer pair G-23 ................................. 153

4-11. Estradiol induced ZP2 in sheepshead minnow by Northern blot ......................... 154

4-12. Induction of sheepshead minnow vitellogenin mRNA following constant aqueous
exposure to xenoestrogens ................................................................................... 155

4-13. Induction of sheepshead minnow plasma vitellogenin following constant aqueous
exposure to xenoestrogens ................................................................................... 156

4-14. Decreased levels of sheepshead minnow vitellogenin mRNA following 16 day
exposure ......................................................................................................... 157

4-15. Decreased levels of sheepshead minnow plasma vitellogenin following 16 day
exposure ......................................................................................................... 158

4-16 Estradiol-depressed transferrin mRNA in sheepshead minnow by Northern
blot ................................................................................................................. 159














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

GENE REGULATION IN TELEOSTS BY ESTRADIOL AND ESTROGEN MIMICS

By

Christopher James Bowman

May 2001


Chairman: Nancy D. Denslow, Ph.D.
Major Department: Pharmacology and Therapeutics

Estradiol plays a critical role in homeostasis, growth, and reproduction. These

physiological effects of estradiol are very tightly regulated at the level of gene

transcription. This regulation is vulnerable to perturbation by environmental

contaminants that mimic estradiol. The biological response over time and with dose by

estrogenic chemicals helps determine the relative risk of hormone-sensitive disease and

cancer in all organisms including humans. This dissertation investigates the effects of

estradiol and environmental estrogen mimics on gene expression in sentinel species.

Estrogen receptor, a nuclear transcription factor, once bound by specific

chemicals binds to specific DNA regulatory sequences to control gene transcription.

Two teleost species, largemouth bass and sheepshead minnows, with unique

characteristics were chosen to evaluate the activation of the estrogen receptor-mediated

pathway. In fish, both vitellogenin and estrogen receptors are known to be induced by

estradiol. In largemouth bass the hypothesis tested was the coordinated induction of

estrogen receptor and vitellogenin mRNA over time after acute exposure to estradiol. An

xii








in vitro method was also developed using liver cells isolated from largemouth bass.

Future studies using this model can be used to evaluate the mechanisms behind the

differential induction of mRNAs.

Sheephead minnow were used to establish and test an in vivo bioassay for

estrogenic chemicals. The dose and time-dependent induction of vitellogenin mRNA and

protein were characterized after exposure to estradiol, ethinylestradiol, diethylstilbestrol,

nonylphenol, methoxychlor, or endosulfan. The decrease of Vtg mRNA and protein

levels induced by estradiol or nonylphenol following transfer to clean water was also

determined.

This research established the baseline experimental responses of estrogen receptor

and vitellogenin mRNAs to estradiol and various estrogen mimics using different routes

of exposure. By differential display additional genes under estrogen control were

identified in these two teleost species, including the zona radiata proteins, transferrin, and

a protein disulfide isomerase, ERp72.













CHAPTER 1
INTRODUCTION AND LITERATURE REVIEW


Background

The Endocrine Disruption Hypothesis

Beginning in 1962 with Rachel Carson's landmark publication, Silent Spring, the

general public became acutely aware that industrial progress did not operate in a bubble,

but held a silent environmental consequence that was growing unchecked (Carson, 1962).

In the years to follow, the United States created the Environmental Protection Agency

and placed new emphasis on pesticide regulation and food safety. Then in 1991, the term
"endocrine disruptor" was coined at the first of a series of scientific work sessions known

as the Wingspread Conferences in Racine, Wisconsin. Simply put, the endocrine

disruption hypothesis is that some chemicals in the environment mimic estrogens and

other hormones, and these can interact with endogenous endocrine systems with

unnatural repercussions.


Supporting Environmental Evidence

Epidemiological studies of humans and observational research of wildlife

following environmental exposure have provided the most realistic evidence for

endocrine disruption. In humans, p,p '-DDE (pp '-dichlorodiphenyldichloroethylene)

concentrations in breast milk have been linked to significantly shortened duration of

lactation (Rogan et al. 1987). Studies of polychlorinated biphenyl (PCB)-contaminated

rice oil in Yusho, Japan and Yu-Cheng, Taiwan demonstrated a unique "fetal PCB

syndrome" in infants of exposed pregnant women (Hsu et al. 1985, Rogan et al. 1988,
1








Yamashita & Hayashi 1985). This syndrome included low birth weight, increased

mortality, rocker bottom heel, and abnormal calcification of the skull among other signs.

In the North American Great Lakes region deficits in neurologic development of children

were positively associated with prenatal exposure of maternally consumed PCB-

contaminated fish. Potentially the most significant impact on human health is on male

reproduction. The issue is two-fold, male reproductive disorders appear to be increasing

across the world and over a large period of time beginning with the introduction of these

environmental contaminants. If this is true, are these environmental contaminants

responsible? Currently under investigation, this issue is an area of tremendous

controversy. Several specific disorders are under scrutiny: cryptorchidism, hypospadias,

testicular cancer, and several parameters concerning poor sperm health (Carlsen et al.

1995, Toppari et al. 1996). None have been conclusively or directly linked to endocrine

disrupting contaminants, but research continues in this area.

Evidence for suspected endocrine disruption is perhaps most convincing in

wildlife. Fish are useful indicators of ecosystem contamination since they are exposed to

all the pollutants in a particular drainage basin due to runoff and erosion. Municipal and

industrial (textile, pulp mill, etc.) effluents have been associated with endocrine

modification and reproductive disruption in fish. These effluents have been reported to

be responsible for testes containing oocytes (intersex) (Harries et al. 1997, Jobling et al.

1998) and increased female-specific proteins in male fish (Folmar et al. 1996, Purdom et

al. 1994). Reduced plasma sex steroids, gonadotropins, egg and gonad size have been

reported in effluents from bleached kraft mills (Van Der Kraak et al. 1992). The Great

Lakes region in North America serves as another example of disruption in wildlife. The

high levels of PCBs, dioxins, and DDT in fish tissue have served as a model of








bioaccumulation and persistence of these contaminants. They have been held collectively

responsible for decreased population in various fish species (lake trout and salmon)

(Guiney et al. 1996, Zabel et al. 1995). These chemicals in the fish have gone up the

food chain to affect bird populations as well. Cormorants, gulls, and terns have been

shown to have the most significant population decline (Allan et al. 1991, Fox et al. 1991,

Giesy et al. 1994), as well as skewed sex ratios in gulls (Hunt et al. 1980). Decreased

larval survival and maturation of English sole in the Puget Sound is another example

(Casillas et al. 1991, Johnson et al. 1988). In Florida, investigators have shown

decreasing population and reproductive abnormalities in alligators from Lake Apopka

(Guillette et al. 1994, Rice et al. 1996, Woodward & Moore 1990).

Due to the scientific, political, and social questions raised by these observations

and ideas, the United States amended the Safe Drinking Water Act (PL 104-182) and the

Food Quality Protection Act (PL 104-170) in 1996 to require testing of potential

xenoestrogens. In addition, because of the complex and unknown nature of how these

chemicals affect biosystems, research into possible mechanisms of action is a priority

(Kavlock et al. 1996). Both mammalian and non-mammalian vertebrates have been used

extensively for in vitro and in vivo laboratory research. Because of the complexity of the

endocrine system there is no single test to determine if a compound will cause adverse

hormonal activity. To complicate matters much of the basic comparative endocrinology

and physiology among relevant species is not well understood, especially in response to

xenobiotic insult.


Supporting Laboratory Evidence

Many laboratory experiments were done to investigate potential mechanisms of

hormonal disruption. Experiments duplicating the known adverse effects of








diethylstilbestrol (DES) in humans showed how sensitive reproductive and

developmental end points are to xenobiotics acting through estrogen mechanisms (Herbst

& Bern 1981, McLachlan et al. 1975, Newbold, 1995). This has opened up the

possibility of other estrogen-mimics having similar or otherwise adverse effects on

humans and wildlife. The effects measured in adults are often only transient, whereas

developing organisms can be permanently imprinted and may be more sensitive to dose;

therefore choosing the appropriate developmental window of exposure is important. It is

very difficult to connect cause and effect even in laboratory experiments since the end

point effects are usually significantly removed in time from the exposure. The

identification of appropriate end points or markers is equally difficult for this reason.

In mammalian models there is good evidence for disruption following

reproductive or developmental exposure to estrogenic chemicals. Reduced seminal

vesicle and prostate weight in rats exposed to op '-DDT (op '-

dichlorodiphenyltrichloroethane) (Gellert et al. 1974), blocked implantation of embryos

by methoxychlor (Cummings, 1990), and changes in territorial behavior of mice exposed

prenatally to op '-DDT, DES, or methoxychlor (vom Saal et al. 1995) are just a few

examples. Sharpe et al. reported a reduction in daily sperm production in rats following

butyl benzyl phthalate and octylphenol (Sharpe et al. 1995). Rats had increased prostate

weights after gestational exposure to very low levels of bisphenol A (Nagel et al. 1997)

which is very similar to the effects exhibited by DES (vom Saal et al. 1997). When

attempting to identify doses in the range of human exposure (Brotons et al. 1995, Olea et

al. 1996), these particular studies are in that range.

In wildlife, studies relevant to environmental exposures have been conducted to

identify potential effects of polluted ecosystems. Some examples are significant








decreases in the rate of testicular growth in rainbow trout following alkylphenol exposure

(Jobling et al. 1996), altered sexual differentiation following alkylphenol exposure

resulting in feminization of male carp (Gimeno et al. 1998b), and altered plasma steroid

hormone levels and poorly organized testes in Lake Apopka alligators (Guillette et al.

1994). Xenobiotic exposure can induce skewed sex ratios in turtles (Bergeron et al.

1994, Crews et al. 1991) and eggshell thinning in birds (Cooke, 1973) as well. Despite

reported examples of endocrine disruption, the mechanisms are still unclear. In the case

of wildlife, there is not enough baseline or basic research to support the necessary

toxicological studies.




Xenobiotics with Estrogenic Activity

One of the most puzzling aspects of xenoestrogens is their common ability to bind

the same receptor despite having very uncommon molecular structures. Some are not

even steroids, such as diethylstilbestrol. Structure-activity relationships to build

predictive models have not been very successful for this reason. The promiscuity of the

estrogen receptor (ER) is well known, but not well understood. It isn't even clear if all

these xenoestrogens even bind the receptor in the same binding site. In general, most of

these xenoestrogens are roughly one thousand times less potent than the endogenous

ligand, E2, using binding affinity for the ER. The only exceptions are the pharmaceutical

estrogens. New classes of compounds that elicit estrogenic activity are still being

discovered. See Figure 1-1 for representative structures of different families of suspected

xenoestrogens.








Pharmaceutical Estrogens

Diethylstilbestrol (DES) and ethinylestradiol (EE2) are two prominent synthetic

chemicals that were custom tailored to function similarly to the endogenous ligand, 170-

estradiol (E2). Ethinylestradiol is currently used in oral contraceptives, and is commonly

prescribed at a dose of 20 to 50 micrograms per day. It also has a slower rate of

metabolism than the endogenous ligand, and much of it is eliminated from the body. So

not too surprisingly, a significant amount of the parent compound (EE2) is released into

the environment in municipal waste effluent (Aherne & Briggs 1989, Desbrow et al.

1998, Tabak et al. 1981). DES, once taken during gestation to prevent miscarriage, is

now known to produce adverse effects and is no longer used (Dieckmann et al. 1953).

The mechanism of action of DES in humans and wildlife is well studied, and continues to

be used as a model compound for xenoestrogen actions (Herbst & Bern 1981, McLachlan

et al. 1975). Both diethylstilbestrol and ethinylestradiol are approximately equipotent to

E2, and have similar mechanisms despite divergent structures.


Organochlorine Compounds

This is largely a class of persistent PCBs and pesticides that are ubiquitous in the

environment. Even though the United States has banned the use of PCBs and DDT since

1977 and 1973 respectively, these compounds are still present in significant amounts in

this country. Concentrations of PCBs, DDT, chlordane and hexachlorocyclohexanes

have been detected worldwide (Iwata et al. 1993) and in rainwater (Rapaport et al. 1985)

which helps explain the enormous distribution. Many of these chemicals are considered

estrogenic by in vitro assays such as the E-Screen (Soto et al. 1994, Soto et al. 1995), or

by in vivo bioassays (Ecobichon & MacKenzie 1974). Some of these organochlorines

such as methoxychlor and o,p '-DDT are considered estrogenic by most screening assays








and have shown mechanisms of action and whole-body effects similar to E2 (Bitman et

al. 1968, Cummings, 1997).


Phenolics

Alkylphenol ethoxylates, 4-alkylphenols, bisphenol A, and some other types of

PCBs fall into this category of proposed xenoestrogens. Alkylphenols are used as

surfactants and emulsifiers in numerous industrial and commercial applications; 360,000

tons were produced worldwide in 1988 (Nimrod & Benson 1996). They are also

components of herbicides, cosmetics, paints, and as spermicides in topical contraceptives.

(Nimrod & Benson 1996) These compounds are thought to be partially responsible for

observed estrogenic effects of municipal and industrial plant effluent, as they have been

measured up to hundreds of microgram per liter (Ahel & Giger 1985, Naylor et al. 1992).

They have been shown by most in vitro and in vivo assays to induce effects similar to E2

(Jobling & Sumpter 1993, Jobling et al. 1996, White et al. 1994). Bisphenol A is a

chemical intermediate for various industrial products including polymers, resins, dyes,

and flame retardants; it is also used in dental sealants and various plastics. Its

estrogenicity is convincing in both in vitro and in vivo assay for both rodents and wildlife

(Brotons et al. 1995, Krishnan et al. 1995, Nagel et al. 1998, Soto et al. 1995, Soto et al.

1997, Sumpter & Jobling 1995). In vivo bioassays have shown the ability of some

phenolic PCBs to exhibit estrogenicity as well, particularly 2',4',6-trichloro-4-biphenylol

and 2',3',4',5'-tetrachloro-4-biphenylol (Jansen et al. 1993, Soto et al. 1995, Young et al.

1995).








Naturally Occurring Compounds

Xenobiotics with potential estrogenic activity do not have to be synthetic as

mentioned above. Naturally occurring compounds in plants as well as fungal metabolites

are foreign to our body, yet some appear to mimic the activity of E2 (e.g. phytoestrogens).

Several of these chemicals are capable of binding to the ER and testing positive in

recombinant-reporter assays (Miksicek 1995, Verdeal et al. 1980). For example,

genistein (flavonoid), coumestrol (flavonoid-derived compound), and zearalenone (fungal

metabolite) all exhibit estrogenic activity as measured by two or more assays. Wood-

derived phytoestrogens, such as beta-sitosterol, could partially account for the proposed

estrogenic activity in pulp and paper mill effluents (Mellanen et al. 1996, MacLatchy &

Van Der Kraak 1995). Even though we know relatively little about these compounds,

there is frequent dietary exposure to humans and infants (soy milk). Therefore, more

research is needed to better understand what dosages have pertinent biological activity in

exposed biota.


Other Compounds

It seems that new groups of chemicals are constantly being recognized as

potential endocrine disrupting chemicals. Most recently the phthalates have been added

to the list of suspected xenoestrogens. These are chemicals used as plasticizers for

polyvinyl chloride and coatings, as cosmetic components, as solvents, or as a leak

detector. Butylbenzylphthalate, butylated hydroxyanisole, and di-n-butylphthalate are the

most interesting. In fact, they seem to bind the ER and positively respond to E-Screen

and gene induction assays, but without in vivo estrogenic responses (Jobling et al. 1995,

Soto et al. 1995). Rather, some evidence suggests they may actually be acting as anti-

androgens in male reproductive assays (Mylchreest et al. 1998). However unlike the








antiandrogen flutamide, phthalates do not seem to block the androgen receptor to elicit

antiandrogenic action (Mylchreest et al. 1999).




Fish Models

The basic features of endocrine-mediated pathways are known to be

evolutionarily conserved in most vertebrates. Despite some unique genes and specific

sensitivities to regulation, the general hormone signaling pathways are thought to be

similar between fish and mammals. When trying to establish cause and effects of

environmental endocrine disruption, fish have been proven to be one of the more affected

species. This diagnosis stems from problems in reproduction and development of

sensitive species in isolated regions. A couple of well known examples include

population declines of lake trout, bloaters, salmon, lake herring, striped bass, and walleye

in the North American Great Lakes region (Guiney et al. 1996, Zabel et al. 1995), as well

as decreased larval survival and maturation of English sole in the Puget Sound site

(Casillas et al. 1991, Johnson et al. 1988). In some Florida lakes, there is evidence for

alligator reproductive impairment (Guillette et al. 1994, Rice et al. 1996, Woodward &

Moore 1990) and some preliminary data in our labs suggests declines in some fish

species, especially largemouth bass, from those same lakes. In general, fish can serve as

relevant biologically-impacted models of environmental health. This is because of their

large numbers, limited range, sensitivity, and their proximity to many potential

environmental exposures. Being aquatic animals, there are various routes of exposure

(gills, food, & dermal) present in our lakes and rivers. These are the same lakes and

rivers that we drink from and dump our wastes into.








Fish can serve as sentinel species illustrating the exposure to and effects of

current levels of pollution. In addition to sentinel species in the wild, fish have also been

used extensively in laboratory studies. The medaka, fathead minnow, goldfish, killifish,

sheepshead minnow, and zebrafish for example are small, easily maintained fish ideal for

large-scale exposure studies. The zebrafish for example serves as a model where the

entire genome has been sequenced. Another larger fish model is the rainbow trout, where

a plethora of background data is available.

For the studies presented in this research a large, environmentally impacted

freshwater fish model (largemouth bass), and a small laboratory estuarine model

(sheepshead minnow) were chosen. One of the major differences is size; larger fish

provide more tissue and material to investigate per sample, allowing for unique

experimental designs such as primary organ culture. Large numbers of small fish on the

other hand are easy to maintain in the laboratory, making realistic exposures and sample

sizes easier to work with. The comparative biology of these two models is also useful

when studying aspects of endocrinology and reproduction. Largemouth bass

reproduction is synchronized and spawns annually, this biological feature enables us to

chart baseline physiology and detect any disruption by xenobiotics. Sheepshead minnow,

on the other hand, is a fractional spawner that serves another purpose in monitoring full

life cycle and generational responses in the laboratory, but within a much shorter period

of time. I would expect that these biological differences result in species-specific

sensitivities to various xenobiotics.

Studies using these model organisms and sentinel species as bioindicators of the

state of the environment are important to detect various biologically-active pollutants and

their potential mechanisms of action. The presence of specific xenobiotics, if not abated,








can continue to accumulate and have the potential to disrupt higher-order organisms such

as humans. The scenario is not new, Rachel Carson predicted such a state over 35 years

ago in her book Silent Spring (Carson, 1962), and this was echoed recently in the context

of endocrine disruption by Theo Colbom in the book Our Stolen Future (Colborn et al.

1997).




Vitellogenin

One of the reasons fish provide such valuable information on environmental

estrogens is the endogenous production of the protein vitellogenin (Vtg). Vtg is a

female-specific yolk precursor protein in oviparous (egg laying) vertebrates (Mommsen

& Walsh 1988). Vtg is a large phospholipoglycoprotein that is thought to be conserved

in function across oviparous vertebrates, but the amino acid sequence is not very well

conserved across species even from the same taxa. Vtg from different species share

small patches of conserved sequences amid larger segments that vary considerably.

Vitellogenesis is initiated through environmental signals and the hypothalamic-pituitary-

gonadal axis. These signals from the brain (gonadotropins) tell the gonads to synthesize

1703-estradiol from testosterone (aromatization). Estradiol is secreted into the blood and

diffuses into the liver. Once in the liver it is bound by estrogen receptors which interact

with chromatin to expose DNA-binding sites. This ultimately induces transcription and

translation of the Vtg gene. Once translated, Vtg undergoes extensive post-translational

modification and is secreted into the blood. It is transported as a dimer and is selectively

taken up by the developing oocyte and specifically cleaved into the egg yolk proteins,

phosvitin and lipovitellin (Figure 1-2) (Mommsen & Walsh 1988).








Model of Hormonal Regulation

Vitellogenin historically has been used as a model for steroid-induction of gene

activation (Tata & Smith 1979). It is keenly specific for estrogens since the 5'-promoter

region of the gene has consensus estrogen response element sequences for ligand-bound

estrogen receptor protein complexes (Burch et al. 1988). Males do not express Vtg at any

appreciable level, but it can be induced artificially, and is thus a useful tool for the study

of cell and molecular biology of gene expression as regulated by E2 (Wahli et al. 1981).

For basic science, it has been used as a tool for characterizing hormone induction profiles

(Bowman et al. 2000, Ryffel, 1978). In fact Vtg mRNA has also been shown to exhibit

differential half-lives in the presence of estrogen, and therefore provides an endogenous

mechanism for characterizing inducible mRNA stability (Brock & Shapiro 1983). There

is evidence to suggest this post-transcriptional mechanism of inducible mRNA stability is

susceptible to disruption by xenoestrogens (Ratnasabapathy et al. 1997).


Biomarker of Estrogenicity

Vitellogenin has been used as a model to answer many questions both in basic

science and for the current topic of environmental endocrine disruption. Vtg is perfectly

suited as an in vivo biomarker of estrogen exposure (Denslow et al. 1999b, Heppell et al.

1995, Palmer & Selcer 1996). The presence of Vtg in male or juvenile oviparous

vertebrates indicates prior estrogen exposure. This is particularly useful in wild

populations of fish present in polluted waters. By quantifying the amount of Vtg in male

fish in vivo, people have identified the presence of an estrogenic component in industrial

and municipal effluents (Folmar et al. 1996, Harries et al. 1999, Purdom et al. 1994).

Currently, it is not clear if there are adverse effects of small amounts of Vtg in males.

There is some evidence that environmentally-relevant levels of Vtg in male flounder can








result in hepatocyte hypertrophy, disruption of spermatogenesis and kidney damage

(Folmar et al. 2001). In addition to plasma Vtg protein, Vtg mRNA can be used to

identify and characterize acute or chronic effects at the molecular level (Bowman et al.

2000, Hemmer et al. 2001). Quantification of Vtg mRNA can provide more sensitive

insight into timing and mechanisms of hormonal disruption.




Estrogen Receptor Mediated Pathway

Steroid hormone-mediated actions are critical for reproduction, the stress

response, growth & development of individuals, homeostasis, and much more. All of the

above events are vulnerable to disruption and/or modulation by xenobiotics. However,

our understanding of the mechanism of such actions is confounded by crosstalk among

different steroids, their respective receptors, organ specificity, cellular environment, and

nonsteroidal signal transduction. Historically, the estrogenic response is the best-

understood steroidogenic pathway of gene activation. This response is thought to be

mostly directed through the ER. There remain many unresolved issues regarding gene

regulation and the dependence on tissue-specific intracellular constituents that dictate the

differential expression of estrogen-responsive genes (Zacharewski, 1998). In

characterizing the estrogenic response one must go from physiological effects seen in

vivo to isolated molecular events that have the capacity to dictate such responses.

The simple ER-mediated pathway starts with the translocation of the lipophilic

ligand (E2) into the cell by simple diffusion (Figure 1-3). Once in the nucleus it binds to

the ER (King & Greene 1984), displacing its chaperone heat shock proteins. On binding

of the ligand, the receptor undergoes a conformational change that allows the formation

of homodimers (Fritsch et al. 1992, Kumar & Chambon 1988) (and possibly








heterodimers) (Glass, 1994). This ligand-bound ER dimer complex then binds to a

specific inverted repeat sequence on the DNA known as the estrogen response element

(ERE) (Kumar & Chambon 1988, Mader et al. 1993). This cis element has been found

near almost all estrogen-regulated genes (Glass, 1994, Lucas & Granner 1992). This

bound complex is thought to recruit the necessary co-factors to initiate transcription of

estrogen-inducible genes (Truss & Beato 1993). Once all the appropriate signals are

present, synthesis of target gene mRNAs begin, followed by the translation of that coding

sequence to active proteins that contribute to the estrogenic response. Known target

genes include: pS2 and cathepsin D in human cells lines (Pilat et al. 1993), c-fos, c-myc,

and c-jun in rats (Weisz & Bresciani 1993), apolipoprotein E in mouse (Srivastava et al.

1997), and vitelline envelope proteins (Larsson et al. 1994), ER, and VTG in fish

(Anderson et al. 1996).

Although the model for this seemingly simple pathway is mostly correct, there are

still a plethora of regulatory factors at each step that are only beginning to be understood.

First is the control of free estrogen or xenoestrogen in the plasma, either through

biosynthesis or differential binding by sex hormone binding globulin (SHBG), a plasma

protein (Joseph, 1994). Then there is the ER itself that has been well characterized, but it

remains unclear how the protein itself is regulated. There is good evidence that the

conformational change of the ER that allows DNA binding is partially ligand-specific

(Fritsch et al. 1992, Paech et al. 1997). Recently, the presence of an ERP isoform was

discovered in rats (Kuiper et al. 1996), and now human, fish and mouse homologues have

been found as well. There is now data supporting how ERoc and P3 work differently

(Paech et al. 1997), and in some cases together (Pace et al. 1997), with multiple ER splice

variants complicating matters (Peterson et al. 1998). Interactions with various orphan








steroid receptors are also beginning to complicate the picture of estrogen regulation

(Giguere et al. 1988). Very recently a third receptor, ERy, identified in ovaries of

Atlantic croaker was discovered (Hawkins et al. 2000). With unique tissue specificity

and phylogenetic diversity, the presence of a third ER promises to raise more questions

than answers.

One of the more contemporary models to explain the diversity of the estrogenic

response is the idea that various cis elements of a particular target gene bind individual

transcription factors and/or steroid receptor complexes. This transactivational complex is

responsible for a particular organ and gene-specific response (Tsai & O'Malley 1994).

There have been several of these proteins identified that act as either co-activators (SRC-

1, GRIP1, AIB1) or as co-repressors of the ER-mediated response (Shibata et al. 1997).

Alternative pathways of ER activation involving membrane receptors have also been

suggested (Loomis & Thomas 2000, Watson et al. 1995). One of the crosstalk scenarios

is epidermal growth factor and its tyrosine kinase receptor activating the signal

transduction cascade to influence the ER-mediated pathway (Ignar-Trowbridge et al.

1992, Kato et al. 1995). Many of these potential pathways involve very specific

phosphorylation of the receptor and how this regulates gene activation (Arnold et al.

1995, Le Goff et al. 1994). Recent evidence suggests the involvement of other pathways,

such as calcium release and mitogen-activated protein kinase activation (Improta et al.

1999, Morley et al. 1992). These types of regulatory features are not well defined yet but

are an area of active interest in steroid research.

Identifying possible mechanisms of how xenoestrogens can alter physiological

events is a complex task. In addition to the gross observations made in the field and the

laboratory following exposure, it is critical to understand the molecular events








responsible behind the scenes. It will be at this level that the interaction of environment

and biology will begin to prove or disprove the endocrine disruption hypothesis. The

following chapters will provide a molecular approach to how two relevant aquatic

organisms respond to estrogen exposure in vivo. The establishment of an in vitro primary

hepatocyte culture system from largemouth bass enables the direct testing of estrogen

stimulation of ER and Vtg mRNA synthesis. This will provide the necessary

information on the basic endocrinology to support future experiments testing the

endocrine disruption hypothesis.





















C

H-C GC3



0

0
o#p!DDT


Hydroxy-PCB


y OH

g-C-CH3 0



0

4-Afkyphenols
OH
Bisphenol A



C02R



C02R
Phthalates


Dlothylstilberol


Genistein


Estradiol-17P

Figure 1-1. Representative chemical structures of different families of
suspected xenoestrogens.


















LIVER


NUCLEUS

$ -RECEPTOR-






1 CYTOPLASM
0-


Figure 1-2.


BLOOD


- ESTROGEN--








VITELLOGENIN


FOLLICLE CELLS


Relationship between the liver and ovary during
vitellogenesis in females. This shows how estrogen induces
synthesis of hepatic vitellogenin which is secreted into the
blood for delivery to the developing oocyte (Adapted from
Tata and Smith, 1979).


OVARY











































Simplified scheme of the estrogen receptor mediated
pathway. Estrogen passively diffuses into the hepatocyte
and is bound by the ER in the nucleus or cytoplasm. Bound
and dimerized receptors bind to specific DNA sequences to
initiate transcription of specific genes such as vitellogenin.
Asterisks indicate possible positions of xenoestrogen
disruption, including other protein-protein interactions in
the nucleus.


-+


Figure 1-3.


F fOXIOXMD3A
UnBIWOURA













CHAPTER 2
COORDINATE IN VIVO GENE EXPRESSION IN LARGEMOUTH BASS
(MICROPTERUS SALMOIDES) AFTER ACUTE ESTRADIOL EXPOSURE


Introduction

If the ultimate goal is to understand and protect ecosystem health, in vivo

experiments are practically impossible to replace with in vitro assays. Even though there

are numerous limitations to in vivo experiments, such as smaller sample size, cost, and

lack of sensitivity, their benefits often outweigh the costs. For example, a single

experiment can provide important information on absorption, distribution, metabolism,

elimination, acute vs. chronic dose, organ weight and morphology, gene and protein

expression, interaction and effects on or with endogenous hormones, mortality, and

reproductive success. Not to mention all the unknown cellular proteins and pathways that

may actually be responsible for the evaluated end points. The basic scientific

understanding of endocrine-active compounds and how they may mimic hormone-

specific responses requires whole-animal investigation.

To investigate estrogen-mediated mechanisms of endocrine disruption, there are

several long-standing hormone-specific assays. In mammalian systems these include

vaginal codification (Allen & Doisy 1923, Edgren, 1994), vaginal epithelial cell

proliferation (Martin & Claringbold 1958), vaginal opening (Allen & Doisy 1924, Edgren

et al. 1966), vaginotrophic response (Folman & Pope 1966), uterotrophic response (Rubin

et al. 1951), uterine glycogen deposition (Bitman & Cecil 1970), and uterine estrogen-

withdrawal bleeding (Schane et al. 1972). For environmentally-impacted animals such as

fish there are few estrogen-relevant end points. One is vitellogenin (Vtg) synthesis in
20








male oviparous species (Folmar et al. 1996). Other assays include altered sex steroids

(Van Der Kraak et al. 1992), organ weights (Munkittrick et al. 1994), and the presence of

ovo-testis by histopathology (Gimeno et al. 1998a, Jobling et al. 1998).


Coordinate Gene Regulation

The integration of signals that ultimately control the estrogenic response is

complex, therefore it is important to characterize the response at the transcriptional and

translational levels to better understand the physiological outcome, while still accounting

for specific gene regulation. Most estrogenic responses are elicited via de novo synthesis

of specific estrogen-regulated genes (Tsai & O'Malley 1994, Carson-Jurica et al. 1990),

including the estrogen receptor (ER) itself in oviparous vertebrates. Characterizing the

transactivation of specific genes following exposure is important to understand the

coordinated tissue and species-specific protein-protein and protein-DNA interactions.

This is partially why gene expression monitored in vivo is so important.

On closer examination of transcriptional events, multiple waves of responses are

seen following a single hormone exposure. Specific tiers of transcriptional activation

exist after a single exposure to a hormone: primary, secondary, and delayed primary

responses (Dean & Sanders 1996). A primary response is defined as those mRNAs that

are up regulated directly by exposure to a hormone. A secondary response is defined as

those mRNAs that are not induced directly by the hormone, but rather by protein

products of the mRNAs originally induced. A delayed primary response refers to mRNA

transcription that results both from a direct interaction of the hormone and additional

factors that are products of the primary response (Figure 2-1). The delayed primary

response would be delayed by several hours relative to the primary response. Induction

of vitellogenin mRNA might be considered a delayed primary response. These types of








varied transcriptional responses have been observed in several different systems and

species (e.g., chicken ovalbumin (Dean & Sanders 1996) and rat a2.-globulin (Chan et al.

1991)). The ordered sequence of transcriptional events is expected to be important for

homeostasis and certainly during reproduction and development. Disruption of these

events is thought to be very sensitive to hormonally active agents.

The estrogenic response in oviparous (egg-laying) animals may represent one

model to study this transcriptional network. The up regulation of ER mRNA and protein

is a primary response, as it is immediate and sensitive following estradiol (E2)-exposure.

The vitellogenin gene found in these animals is thought to elicit both a small primary

response and a larger delayed primary response (Pakdel et al. 1990). It is believed that

the additional ER synthesized in the primary response is responsible for the delayed

response seen with Vtg (Pakdel et al. 1991). A secondary response is seen with genes

that do not directly bind intracellular receptors, but instead are activated or induced by

one of many potential intermediary proteins (Dean & Sanders 1996). These intermediary

proteins could consist of transcription factors, kinases, mRNA stability proteins, or

intracellular cofactors. The classic example of this tiered response was first shown by a

steroid (20-hydroxyecdysone)-induced transcription factor that induced waves of

polytene chromosome puffing in Drosophilia (Thummel, 1995). At this time, the

evidence for estrogenic secondary response genes that are a direct effect of these

intermediary proteins is scarce, but remains an area of active research. Part of the

problem with studying estrogen target genes across species is that very few have been

discovered. Most of what is understood concerning E2 is its proliferative effects in

reproductive tissues and its signal to synthesize vitellogenin (precursor to egg yolk








protein). Because of this, Vtg in frogs, chickens, fish, and birds have been used as model

systems to study hormone induction (Ryffel, 1978).

By studying the regulation of ER and Vtg gene induction following E2-exposure it

is possible to learn more about the establishment of complex gene networks in a system.

In addition to understanding how known genes are coordinately regulated, the discovery

of novel or previously unrecognized estrogen-regulated genes may lend insight into the

dynamics of gene networks. One mechanism for identifying such unknown estrogen-

regulated molecules is differential display (Liang & Pardee 1992). Samples collected

over time following an acute exposure, once analyzed by differential display, could lead

to such discoveries.


Largemouth Bass

Largemouth bass (Micropterus salmoides) (Figure 2-2) is a native freshwater fish

found in the Great Lakes region and the Mississippi River Basin from southern Quebec

down to the Gulf of Mexico. In addition to its large natural distribution, it is stocked in

over 30 states as a game fish. Because it is a piscivorous predator and can grow quite

large, it is usually at the top of the fish food web in its natural territories. This capacity to

embody much of the freshwater fish population and its suspected sensitivity makes it a

model sentinel species. In recent years fish biologists and researchers have documented

declining numbers in local Florida lakes. In addition, stocked fish reproduction is now

suffering in reclaimed wetlands due to unknown circumstances in the environment. One

hypothesis is the residual presence of agricultural chemicals in these lands. Because

some of the old, persistent agricultural chemicals have demonstrated estrogenic potential,

this is a plausible mechanism of endocrine and reproductive dysfunction.








Experimental Objectives and Hypothesis

The objective of this set of studies was to prepare gene-specific cDNA probes to

characterize the coordinate regulation of ER and Vtg mRNA relative to plasma Vtg in

largemouth bass (LMB). Another priority was to find other estrogen-regulated genes to

complement the understanding of coordinate gene expression in fish. The primary

hypothesis tested was that E2 would induce a primary and delayed primary hepatic

transcriptional responses over time for ER and Vtg mRNAs, respectively. This acute

exposure to E2 would also elicit a cascade of differential responses at the mRNA level

through up- and down- regulation of known and unknown genes that could be identified

by differential display.




Materials and Methods


Fish Collection and Maintenance

Largemouth bass (Micropterus salmoides) were purchased from American

Sportfish Hatchery (Montgomery, Alabama). They were either maintained at the Aquatic

Toxicology Facility Lab at the University of Florida under the direction of Dr. Evan

Gallagher or at the United States Geological Survey- Caribbean Science Center under the

direction of Dr. Timothy Gross. All experiments were designed and conducted under my

direction. Fish were acclimated in aerated 104 to 250 gallon fiberglass tanks (Figure 2-

3A and 2-3B), under constant conditions of 21 +/-2' C for a minimum of one week prior

to exposures. Dissolved oxygen, total ammonia content, and pH were monitored, but did

not vary significantly during acclimation and exposure periods. The fish were exposed to

ambient light concentrations, and fed Purina Aquamax 5D05 fish feed (St. Louis, MO)








once a day. Many lab personnel assisted with the sample collection (Figure 2-3C and 2-

3D).


Experimental Exposures

All three experiments were single acute exposures administered by intraperitoneal

injection. In the first experiment, adult (> 1 year old) male LMB were injected with -2

mg 17P-estradiol (E2)/Kg dissolved in dimethyl sulfoxide (DMSO). The fish were

returned to the tank, and sampled at 0, 6, 12, 24, and 48 h post injection. Separate control

fish were injected with DMSO or not injected at all. These fish were sampled in July of

1998 and had low to no observable spermatogenic activity (Figure 2-4A).

In the second experiment, adult male LMB were injected with varying doses of E2

or estrogen mimics, followed by sample collection 48 h later. Chemicals and doses used

were as follows: E2 (0.0005, 0.005, 0.05, 0.5, 5.0 mg/Kg), ethinylestradiol (EE2) (0.005,

0.05. 0.5 mg/Kg), nonylphenol (NP) (0.05, 0.5, 5.0 mg/Kg), methoxychlor (MXC) (0.05,

0.5, 5.0 mg/Kg) or op -dichlorodiphenyl-trichloroethane (op '-DDT) (0.05, 0.5, 5.0

mg/Kg). All chemicals were dissolved and diluted in DMSO. Control fish were injected

with DMSO or not injected at all. NP, MXC, and o,p '-DDT were purchased from

ChemService (West Chester, PA). These fish were sampled in November, 1999 and had

low to moderate spermatogenic activity (Figure 2-4B).

In the last experiment, 100 juvenile LMB (< 1 year old) were injected with a -2

mg E2/Kg. The fish were returned to the tank, and sampled at 0.25, 1, 2, 4, 7, 14, or 21

days post injection. Separate control fish were injected with DMSO or not injected at all.

Although these fish were less than one year old, the gonads of these fish were developing

at the time of sample collection. Following histological examination of the gonads, only








fish exhibiting male characteristics were analyzed (Figure 2-4). These fish were sampled

in January, 2001 and had moderate to high spermatogenic activity (Figure 2-4C).

Blood and livers were collected from all experimental and control fish. Blood

was drawn using a Icc tuberculin syringe with 20 guage needle. Blood was stored in a

heparinized vessel with aprotinin at 4C until separation. After centrifugation for 20 min

at 4800 x g the plasma was aspirated and stored in aliquots at -70*C until analyzed. One

aliquot of the plasma was sent to the lab of Dr. Tim Gross, where steroid analyses by

radioimmunoassay were conducted by his staff. The liver was excised, then immediately

flash-frozen in liquid nitrogen and stored at -70C until analysis. Gonads were collected

from all fish and preserved in 10% buffered formalin until paraffin-embedding and

sectioning for histological examination for sex determination. Paraffin embedding, tissue

sectioning and H&E staining were performed by the Histology Core facility at the

University of Florida.


RNA Isolation

RNA was isolated as described previously by one of two methods (Bowman &

Denslow 1999). Briefly, the individual liver tissues were processed for RNA using the

acid phenol guanidinium-isothiocyanate (Chomczynski & Sacchi 1987), or using RNeasy

kits from Qiagen (Valencia, CA). The samples were treated with Proteinase-K, then

measured at 260 and 280 nm using a spectrophotometer. RNA samples used in

differential display were also DNase treated. The 260 nm measurement was used to

estimate the concentration of total RNA recovered from the isolation. The 260/280 ratio,

as well as a 1% agarose-formaldehyde gel stained with ethidium bromide, were used to

verify the quality of the RNA in each sample.








Estrogen Receptor and Vitellogenin RT-PCR

Because of the high homology of ER across vertebrate species, nucleotide specific

primers were designed based on sequence homologies in the hormone binding domain of

human, chicken, frog, and trout ER a. Vitellogenin, however, had very disparate

nucleotide similarities, so degenerate primers were designed based on protein sequence

homologies in the 3'-end of the Vtg sequences available at that time. The ER a primer

sequences were: upper -TCA CCA TGA TGA CCC TGC TCA; lower -TGC TCC ATG

CCT TTG TTG CTC. The sequences of the Vtg primers were: upper -CAR GTN YTN

GCN CAR GAY TG; lower -GCA YTC NSW NGC RTC NCK RC. Both sets of primers

were designed used Oligo 5.0 (Cascade, CO).

Oligo-dT primers, dNTPs, 5X transcription buffer (Life Technologies; Rockville,

MD), and Superscript II (Life Technologies) were used to reverse transcribe 2 gg total

RNA. Specific ER (10 pmol/gl) or degenerate Vtg primers (80 pmol/gl) in lOX reaction

buffer with MgC12 (Perkin Elmer; Foster City, CA), dNTPs, AmpliTaq (Perkin-Elmer),

and 2 gl of cDNA from the reverse transcription reaction were used to amplify portions

of the ER or Vtg gene by the polymerase chain reaction (PCR). The PCR conditions for

both genes were: hold at 80C for 3 min; hold at 94C for 3 min; 35 cycles of 940C for

45sec, 52C for 90sec, 72C for 45sec; hold at 72C for 10 min; and hold at 40C for one

hour. The PCR products were analyzed and purified by 1.2% agarose gel electrophoresis

(Figure 2-5).


Cloning and Sequencing of Isolated cDNAs

The single bands corresponding to the amplified ER or Vtg cDNA were extracted

from the gel and purified using Qiaquick gel extraction spin columns (Qiagen). The








purified PCR products were ligated into a pGEM-T Easy vector (Promega; Madison, WI)

(Figure 2-6) and transformed into E.coli (DH5a). Plasmid clones containing the

amplified fragments were picked randomly and purified for sequence determination and

probe preparation. Using an ABI PRISM Dye Terminator Cycle Sequencing Kit (Perkin

Elmer) and M13 primers from the DNA Synthesis Core Facility (Biotechnology Program,

University of Florida), sequencing reactions were performed on the isolated plasmid

preparations, and submitted to the DNA Sequencing Core Facility at the University of

Florida for sequence determination. BLAST (Altschul et al. 1997) and multiple sequence

alignment programs (Corpet, 1988) were used to analyze the results.


mRNA Quantification

Complementary DNA probes were made using the cloned ER or Vtg fragments.

Templates for each probe were cut from the plasmid vector using EcoR (Figure 2-6).

The digestion reactions were purified by gel electrophoresis and extracted using Qiaquick

spin columns (Qiagen). These templates were used for the synthesis of [(X-32p]-labeled

cDNA probes using a Strip-EZ DNA Kit (Ambion; Austin, TX) according to

manufacturer's instructions, and purified using TE-Midi SELECT-D, G50 spin columns

(5 Prime-3 Prime, Inc.; Boulder, CO). A O-actin cDNA probe (GenBank Accession #)

was also prepared as described for ER.

ER and Vtg mRNA were quantified as described previously (Bowman &

Denslow 1999, Bowman et al. 2000). Briefly, for Northern blot analysis, 12 g.g of total

LMB liver RNA was denatured and separated on a 1% agarose-formaldehyde gel. The

RNA was transferred to a nylon membrane (Biodyne B, Life Technologies) using

downward capillary action, followed by UV-crosslinking using a Stratalinker 1800

(Stratagene; La Jolla, CA). Nylon membranes were stained with methylene blue to verify








successful transfer and even loading of the samples (Herrin & Schmidt 1988). The

membranes were pre-hybridized in a glass cylinder using a Techne Hybridiser oven with

ExpressHyb hybridization buffer (Clontech; Palo Alto, CA) for 30 min at 68C. The

membranes were then incubated at 68C for one hour in fresh hybridization buffer

containing approximately 3 x 105 dpm ER, Vtg, ERp72 or P-actin probe/mL. Nylon

membranes were then washed twice with 2X SSC, 0.1% SDS for 20 min at 250C, then

twice with 0.1X SSC, 0.1% SDS for 30 min at 60'C or 68C. The nylons were then

wrapped in saran wrap and exposed to BioMax MR X-ray film (Eastman Kodak;

Rochester, NY) for visualization, and were exposed to a phosphorscreen for image

quantification using a Phosphorlmager (Molecular Dynamics, Inc.; Sunnyvale, CA).

For the quantification of ER and Vtg mRNA, sense ER and Vtg cRNA was

transcribed in vitro from the vector (Figure 2-6) using a Megascript Kit (Ambion)

according to the manufacturer's suggestions. Sense ER and Vtg cRNA was analyzed by

gel electrophoresis and quantified by spectrophotometry. Concentrations from 0.01 ng to

5 gig Vtg cRNA serving as standards were denatured and loaded onto a Biodyne B nylon

membrane using a slot blot apparatus (Schleicher and Schuell; Keene, NH). Twelve jtg

of sample were denatured in denaturing buffer (containing 20X SSC, formamide, and

formaldehyde) and loaded into slots on the same membrane. The nylon membrane was

stained with methlyene blue and hybridized as described above for Northern blots (Figure

2-7A and 2-7B).

To standardize gene expression across individuals, Northern and slot blots were

stripped and re-probed with P-actin. All ER and Vtg mRNA sample values obtained

were corrected individually to P-actin levels. Data was collected using a Phosphorlmager








as described above. Actual ER and Vtg mRNA values reported were calculated from a

standard curve generated from the synthesized cRNA standards (Figure 2-7C).


Differential Display RT-PCR

Differential display-reverse transcription polymerase chain reaction (RT PCR)

was performed with the RNAimage mRNA Differential Display system (GenHunter;

Nashville, TN) using one-base anchored oligo-dT primers (Liang et al. 1994). DNase-

treated total RNA (0.2 gig), isolated from control or treated LMB livers was reverse

transcribed using 0.2 jtM of anchor primer and 100 U MMLV reverse transcriptase in a

total volume of 20 jiL as described by the manufacturer. For each condition, we used

three separate liver samples to distinguish false positives. PCR reactions (20 IiL) were

performed following the RNAimage protocol and included one-tenth volume of the

reverse transcription reaction, 10 mM Tris-HCl (pH 8.4), 50 mM KC1, 1.5 mM MgC12,

0.2 jIM anchor primer and arbitrary primer, 2 gtM dNTPs, 2.5 piCi-[33P]-dATP (2000 to

4000 Ci/mmol), and 1 U AmpliTaq DNA polymerase. Primer pairs used are indicated

on each figure. After an initial denaturing step of 940C for 5 min, 40 PCR cycles were

run with the following conditions: 940C for 15 sec; 40'C for 2 min; 72C for 30 sec;

followed by a 10 minute 72C extension step. Aliquots of each PCR reaction were heated

for 3 min at 80*C with DNA sequencing loading dye and separated by electrophoresis on

5% denaturing Long Ranger gels. The gels were dried under vacuum at 800C and

exposed to Biomax MR X-ray film for 18 to 48 h.

Bands of interest were located on the gel, cut out, and then soaked in 100 PiL of

ddH20 for 10 min, followed by boiling for 20 min to elute the DNA. A fraction of the

material was used as template with the same primer pairs to reamplify the DNA. A 40

jLL reamplification reaction was prepared for each cDNA with arbitrary primer 10, 21, or








23 (2 tM) and anchor G- or C-Tl primer (2 jtM). The PCR conditions were set up as

follows: 80C for 3 min 95C for 2 min -, 40 cycles of 94C for 15 sec, 40'C for 2

min, 72C for 30 sec -+ 72C for 10 min -+ 4C soak. Twenty microliters of the PCR

samples were run on a 1% agarose gel and stained with ethidium bromide. If the size of

the reamplified product was correct, the cDNA band was gel purified, cloned and

sequenced as described above.


Protein Analysis

The presence of Vtg in serum samples was verified by electrophoresis using Tris-

tricine gels and Western blot analysis as described previously (Denslow et al. 1997a). All

plasma samples were diluted 50-fold in IX Laemmli sample buffer and 10 ptL were

applied to separate wells of 7.5 % polyacrylamide gels (0.2 JIL plasma per well).

Samples of purified LMB Vtg (0.1 and 0.5 jtg) were included in separate wells as

positive controls. Multimark molecular weight markers (Novex; San Diego, CA) were

used to calibrate the gels. The gels was electroblotted to PVDF membranes (Immobilon-

P-Millipore; Bedford, MA) in 10 mM MES (morpholino ethane sulfonic acid), 10%

methanol, 0.01% SDS, pH 6.0, overnight at 20 V and 4C. For Western blot detection,

the membrane was blocked with 5% nonfat dry milk in TBSTZ (10 mM Tris, 150 mM

NaCl, 0.05% Tween, 0.02% sodium azide, pH 7.2) at 25C for 2 h. The blot was

incubated with primary monoclonal antibody 1C8 HL 1364 in blocking solution,

developed with a secondary goat-anti-mouse alkaline phosphatase-linked antibody

(Pierce; Rockford, IL) and developed with bromochloroindoyl phosphate/nitro blue

tetrazolium as described previously (Denslow et al. 1997b).








A direct ELISA (enzyme-linked immunosorbent assay) was used to quantify

plasma Vtg as described previously (Denslow et al. 1999b, Folmar et al. 1996, Folmar et

al. 2000) using avidin-biotin complex reagents (Pierce). The overall sensitivity of the

ELISA used in this study is 2 jig/mL for plasma Vtg. The assay itself is more sensitive by

a factor of 200 when pure Vtg is used. We are able to measure 0.5 ng Vtg per ELISA

well in a volume of 50 gtL. The lower sensitivity for plasma samples is due to the

requirement to dilute plasma 1:200 to eliminate interferences. The linear portion of the

standard curve extends from 0.01 to 0.8 jtg/mL purified Vtg. All of our standard curves

are prepared with added control male plasma, diluted to the same concentration as the

samples, so that the samples and standards have the same composition. Two different

standard curves were performed with male plasma diluted to 1:200 (for the samples

containing the least amount of Vtg) and 1:10,000 (for the highly induced samples). On

occasion a sample will have a Vtg concentration that is outside these standard curves. In

that event, we repeat the assay with the appropriate standard curve. The primary

antibody used was 3G2 109AB raised against striped bass Vtg. The plate coated with

samples and standards was incubated with the monoclonal antibody overnight at 41C in a

humidified chamber. For the rest of the assay the following reagents from Pierce were

used: goat-anti-mouse IgG (H&L), biotin, and the streptavidin alkaline phosphatase

conjugate. After washing the plate, Vtg was quantified colorimetrically with the alkaline

phosphatase substrate, p-nitro phenyl phosphate (in carbonate buffer with 2 mM MgC12,

pH 9.6) at 405 rum in an ELISA plate reader. All samples and standards were run in

triplicate. The coefficients of variation and correlation coefficients for this assay were

<10% and > 0.95% respectively.







Results


Estrogen Receptor and Vitellogenin Sequencing and Characterization

Because the cloned LMB ER and Vtg fragments were initially obtained by PCR,

five independent clones in both orientations (both strands) were sequenced. Sequence

analysis revealed distinct sequences for the two genes. A BLAST search (Altschul et al.

1997) of each showed specific homology to most of the ER a or Vtg sequences in the

database. As shown in Figure 2-8, the LMB ER ax sequence (Genbank accession number

AF253062) was aligned with ER ax sequences from human to fish. The LMB Vtg

sequence (Genbank accession number AF169287) had greater than 70% identity to the

various Vtg sequences (Figure 2-9). The Vtg sequence is not as well conserved relative

to the ER so there is less homology across species for this gene.

In order to verify that the ER and Vtg mRNAs were induced by E2 as expected,

ER and Vtg gene expression were analyzed by Northern blot. Using random-primed

cDNA probes from these cloned cDNAs by Northern both genes were up regulated in

response to estrogen (Figure 2-10A). ER mRNA in LMB is approximately 3.5 kb in

length. The ER cDNA probe used in these studies was shown to be specific for ER a

without crossreactivity to ER 03 in LMB (unpublished observations). Our Vtg probe

detected 3 distinct bands at 5.0, 3.3, and 1.7 kb for this mRNA in exposed samples only

(Figure 2-10A). Crossreactivity of the Vtg probe with 28S (4.0 kb) and 18S (1.9 kb)

rRNA does not account for the smaller two Vtg bands since control samples do not

present the bands (Figure 2-1OA and 2-10C). All lanes were loaded with equal amounts

of total RNA. Furthermore, the calculated sizes of the Vtg bands are different than the

rRNA bands. Following hybridization with ER a and Vtg, a P-actin probe showed a








steady amount of P3-actin mRNA (2.1 kb) across both control and exposed sets of RNA

samples (Figure 2-1OB). RNA size markers (Promega), 28S rRNA (4.0 kb), and 18S

rRNA (1.9 kb) were used to calibrate the blots when determining the respective ER and

Vtg mRNA sizes.


Regulation of mRNAs in Adult Largemouth Bass 48 Hours Post Injection

Following the injection of E2 (2 mg/Kg), plasma steroids were measured at 0, 6,

12, 24, and 48 h (Figure 2-11). Plasma testosterone stayed below 500 pg/mL for the

duration of the experiment and was not significantly different from controls at any of the

timepoints. As expected, plasma estrogen increased approximately 6-fold over controls

by 6 h post injection. The average amount measured at 6 h, 2,806 pg/mL, did not vary

significantly out to 48 h post injection. The average body weight of these adult fish were

165 g.

The expression of ER and Vtg mRNA was analyzed by both Northern blot

(Figure 2-12A and 2-12B) and slot blot (Figure 2-12D and 2-12E). Since slot blot values

were obtained using a standard curve and are corrected for individual P-actin values,

these were considered more quantitative. Both methods demonstrated the same trends,

ER and Vtg mRNAs were detectable by 6 h and continued to increase for the duration of

the experiment (Figure 2-12). There was a distinct increase in ER mRNA by 12 h, and

by 48 h there was approximately 3.68 pg/pg total RNA. Following injection, Vtg mRNA

nearly tripled by 12 h and had greater than a 10-fold increase by 48 h (approximately 533

pg/gg total RNA). Estrogen receptor mRNA levels appeared to increase slightly faster

than Vtg mRNA, particularly evident by 12 h post injection (Figure 2-12A and 2-12B).

Although not seen by Northern, the slot blot data detected a low basal level of Vtg

mRNA (0.04 ng/lig total RNA) in control fish.








Plasma Vtg, analyzed by both Western blot and ELISA, demonstrated the

expected timed induction relative to Vtg mRNA. In Western blots, two bands

corresponding to 160 and 180 kDa were seen within 6 h post injection. The 160 kDa

band appeared to be more highly expressed and more consistent at 12 and 24 h. By forty-

eight hours, both bands plus some intermediate size fragments, presumably breakdown

products of the 180 kDa band, are visible.

Determination of plasma Vtg by ELISA is more quantitative than by Western

blot. Standard curves using purified LMB Vtg coupled with control male plasma were

used to generate standard curves. Following injection, Vtg was measurably induced by 6

h, then increased up to 1 mg/mL at 24 h. By forty-eight hours Vtg protein accumulated

in the plasma to approximately 4 mg/mL (Figure 2-13).

In order to look for other genes that might also be regulated by estrogens,

differential message display RT-PCR was used over time following acute exposure

(Figures 2-14 and 2-15). Equal amounts of control or exposed samples were loaded on

the gel for side-by-side comparison. As expected some mRNAs were induced by

estradiol in the same time frame as Vtg. Others were induced earlier; some even

returning to baseline by 48 h post injection. PCR-amplified products using primer pairs

G-23, C-l, and G-10 were a good illustration of some differentially regulated genes. The

arrows on the left of Figure 2-14 show examples of specific mRNAs that are down

regulated by E2. As shown by the arrows on the right side of the gel, some cDNAs were

even transiently regulated, up by 12 h and down by 48 h (Figure 2-14). As would be

expected, the regulation of some cDNAs did not change over time following E2-exposure

as indicated by the arrows on the left in Figure 2-15. Arrows on the right illustrate that

some cDNAs were up regulated post injection (Figure 2-15). Twelve of the bands shown








were cut out of the gel, cloned, and sequenced in an attempt to identify what mRNA they

represented. The sizes of the amplified products were from 75 to 560 bp in length. Since

all the cDNAs are amplified from their 3'-ends, sequence information often does not get

into the coding region making gene identification difficult.

One of the few bands identified so far is Vtg (Figure 2-15). The other is a protein

disulfide isomerase-related protein, known as ERp72. This identity was discovered

following BLAST analysis and aligning the predicted amino acid sequence with that of

known human, rat, and mouse (Figure 2-16A). Verification of this E2 up regulation over

time by Northern is shown in Figure 2-16B and 2-16D. Largemouth bass ERp72 mRNA

appears to be the first protein disulfide isomerase-related protein observed in teleosts and

is approximately 2.6 kb in size. Following estradiol-injection, LMB ERp72 mRNA

reached maximum levels (approximately 4-fold) by 24 h, and rapidly decreased by 48 h,

but did not return to near basal levels until 21 days post injection (Figure 2-16D). This

pattern of induction over the first 48 h is different from that observed on the DD-RTPCR

gel (Figure 2-15), demonstrating the importance of validating all observed responses.

One possible reason for this discrepancy is that the clone sequenced from the band on the

gel is only one of the many cDNAs that may be present at that particular size, and may

not represent the higher intensity band seen on the gel (Figure 2-15).


Plasma Vitellogenin Dose Response

In the second experiment plasma sex steroids and Vtg were measured from all

male fish (Figure 2-17 and 2-18). These measurements were taken forty-eight hours

following the injection of several different doses and chemicals. Forty-eight hours

following the 5 mg E2/Kg dose, plasma E2 increased approximately 4-fold. Although

0.05 and 0.5 mg E2/Kg did not increase plasma E2, the two lowest doses (0.0005 and








0.005 mg/Kg) increased plasma E2 levels significantly over controls. Nonylphenol

increased plasma E2 in a dose dependent manner up to 600 pg/mL (Figure 2-17A). Both

of these higher doses of NP increased plasma E2 significantly over controls (P < 0.05).

All doses of MXC significantly increased plasma E2 up to 500 pg/mL by 48 h post

injection (P > 0.05). Plasma estradiol levels were not affected by EE2 or DDT. Plasma

1 1-ketotestosterone (1 1-KT) was highly variable but was not statistically different from

controls between any of the chemicals or dose, with the exception of the 0.05 and 5

mg/Kg of E2 (Figure 2-17B). Plasma steroids in placebo-treated animals (DMSO) taken

at 48 h did not differ from uninjected control fish (data not shown).

Estradiol induced plasma Vtg in a dose-dependent manner up to approximately 9

mg/mL by 48 h post injection (Figure 2-18A). The two lowest doses of E2 (0.0005 and

0.005 mg/Kg) did not appear to induce plasma Vtg. Therefore the lowest dose with an

observed effect on Vtg induction was 0.05 mg E2/Kg. Acute ethinylestradiol-exposure

also elicited a dose dependent increase in plasma Vtg (Figure 2-18B). All doses of EE2

tested increased plasma Vtg; the lowest dose induced plasma Vtg accumulation up to 4

mg/mL. As shown in Table 2-1, none of the doses of NP, MXC, or DDT induced plasma

Vtg over the limit of detection by 48 h post injection.


mRNA Characterization After Acute Exposure in Juvenile Largemouth Bass

In the third experiment, samples were collected at 0.25, 1, 2, 4, 7, 14, and 21 days

following the injection of-2 mg estradiol/Kg. The average body weight of these fish

was 42.2 g. Sex steroids of these fish were not analyzed because of insufficient

quantities of plasma. Analysis of ER and Vtg mRNA by Northern blot was for

qualitative purposes (Figure 2-19A and 2-19B). Since slot blot values were obtained

using a standard curve and were corrected for individual P-actin values, these were








considered more quantitative. Both methods show the same pattern (Figure 2-19). Both

ER and Vtg mRNA were up regulated by 6 h (time point 0.25). The pattern of induction

out to 2 days was similar to that observed in the earlier experiment with the adult LMB,

quickly increasing to maximal levels. By day 7 and for the duration of the three weeks,

both ER and Vtg mRNA returned to levels observed for the 6 hour time point. In these

juvenile fish, ER mRNA was maximally induced by 2 days post injection with -3.5

pg/gg total RNA. Vitellogenin mRNA was induced up to -138 pg/lg total RNA,

approximately 30-fold over the basal level. Estrogen receptor mRNA appeared to

increase faster over time than Vtg in the initial 6 h, and by four days return to minimal

plateau level quicker as well (Figure 2-19). Although not apparent by Northern, the slot

blot data detected a low basal level of ER mRNA (1 pg/pig total RNA) and of Vtg mRNA

(0.004 ng/pgg total RNA) in control fish.

Plasma Vtg, analyzed by ELISA, demonstrated the expected timed induction

relative to Vtg mRNA. Following injection, Vtg was measurably induced by 6 h, then up

to 0.48 mg/mL at 24 h, and by 2 days it accumulated in the plasma to approximately 1.4

mg/mL (Figure 2-20). Rather surprisingly, by 4 days plasma Vtg levels had already

begun to decrease, going down to 0.76 mg/mL. At 21 days it was still detectable in the

plasma at levels around 0.13 mg/mL.




Discussion

Unfortunately, the same features that make LM.B a model sentinel species also

make it difficult to study in the laboratory. Specifically, its large size and dominant

presence in freshwater lakes make it a representative species in these ecosystems. Its

large size and apparent sensitivity to environmental perturbation also make it hard to








raise and maintain in tanks, relative to the care of smaller laboratory model fish species.

LMB reproduction is synchronized over the course of each year. Thus, it is useful for

monitoring reproductive outcomes, but makes it difficult for short-term laboratory studies

because the outcome my depend on the time of the year. To monitor this possible

confounding effect, the fish from each experiment were examined by histology to

determine reproductive status during each experiment (Figure 2-4). It was not expected

that the exposure end points measured in the male fish from all these experiments would

be affected by reproductive status.


Largemouth Bass Estrogen Receptor and Vitellogenin

To begin these studies, the first step was to clone portions of the ER and Vtg gene

from LMB. Since the amino acid sequences of ER are similar across species; human, rat,

frog, and trout ER sequences were used to design primers. Because the DNA-binding

domain is so similar across all nuclear receptors, the primers were made to a homologous

region specific for only ER a using the hormone-binding domain. Based on the

sequences used to design these primers the predicted size of the amplified product would

be approximately 550 bp. Using control and estrogen-treated LMB a band corresponding

to this size was obtained by PCR (Figure 2-5A).

The predicted amino acid sequence from the cloned LMB ER cDNA fragment

lined up with greater than 80% identity with the published sequences for rainbow trout

(Pakdel et al. 1990), gilthead seabream (Socorro et al. 2000), and tilapia (Tan et al. 1996).

It also has significant similarity ( > 65%) to the human (Greene et al. 1986), rat (Koike et

al. 1987), and mouse (White et al. 1987) ER a (Figure 2-8). At the time this sequence

was cloned, there was only one estrogen receptor identified in fish. Therefore all the

work presented in this dissertation is of ER a only. In recent years an ER 13 was








identified in teleosts (Tchoudakova et al. 1999, Xia et al. 1999). Just last year there was

even a report of a third isotype, ERy, in Atlantic croaker (Hawkins et al. 2000). The

unique tissue and ligand specificities of all three isoforms have implications for the

research presented in this dissertation and are areas of future study in the Denslow

laboratory. Currently both ER P and y isoforms have been cloned and sequenced for

LMB (Sabo-Attwood et al. in preparation).

The amino acid sequence of vitellogenin is poorly conserved across oviparous

species. When the available sequences in Genbank are aligned, sequence similarity exists

only in short segments of amino acids along the full length of the protein with the

exception of the phosvitin region, where it is greater than 50% phosphoserine. When

designing our Vtg primers, a region close to the C-terminus of the protein was selected

where there was the most apparent conservation among sequences for trout, sturgeon,

lamprey, and two mummichog sequences. These primers are generally useful for cloning

Vtg mRNA segments from a large variety of fish including largemouth bass, Genbank

accession number AF169287 (Bowman & Denslow 1999), gilthead seabream, Sparus

aurata (Funkenstein et al. 2000), bluegill, tilapia, koi, trout, and pinfish (Bowman and

Denslow, unpublished).

The LMB Vtg cDNA fragment cloned was from the 3' region of the very large

and heterogeneous Vtg mRNA (5 kb estimated). The predicted amino acid sequence for

this LMB Vtg cDNA fragment had greater than 70% homology to the other published

fish sequences, specifically to the mummichog (LaFleur et al. 1995a), tilapia (in Genbank

only, accession number AF017250), and trout (Mouchel et al. 1996). Chicken (van het

Schip et al. 1987), and frog (Gerber-Huber et al. 1987) Vtg sequences had approximately

40% homology (Figure 2-9). There is evidence to suggest the presence of two Vtgs in








teleosts (Bowman et al. 2000, LaFleur et al. 1995b, Matsubara et al. 1999). The two Vtgs

in flounder have been reported to have distinct functions in egg maturation (Matsubara et

al. 1999). In the past month the Denslow lab isolated a fragment of a second Vtg in

LMB.

Since ER and Vtg mRNAs had not been previously characterized in LMB, total

RNA from liver was subjected to Northern blot hybridization using gene-specific probes.

Northern blot analysis of ER mRNA revealed one band of approximately 3.5 kb only in

the E2-treated fish (Figure 2-10A). This is the same size as the primary liver ER mRNA

(3.5 kb) reported in E2-treated rainbow trout (Pakdel et al. 1989). Additional transcript

sizes of ER have been reported in trout liver (4.5 kb) (Pakdel et al. 1989), trout pituitary

(1.4 kb) (Pakdel et al. 1990), and catfish liver (several ER a sizes between 1.5 and 10

kb)(Patino et al. 2000), but these were not evident in LMB liver using this cDNA probe.

Northern blot analysis of LMB Vtg mRNA revealed three bands of 5.0, 3.3, and

1.7 kb in length in E2-exposed fish only (Figure 2-10A). The most intense band at 5.0 kb

is of sufficient size to contain the entire predicted coding sequence for LMB Vtg (160 to

180 kDa) (Figure 2-13A). Largemouth bass Vtg appears similar in size to that reported in

sheepshead minnow (5.0 kb) (Bowman et al. 2000), but smaller than the reported size for

rainbow trout (6.6 kb) (Le Guellec et al. 1988) and tilapia (6.5 kb) (Lim et al. 1991). Size

calibration techniques, however, may account for some of the apparent species difference

in Vtg mRNA sizes. The two minor bands observed in LMB (3.3 and 1.7 kb) are too

small to contain the entire predicted coding sequence for Vtg. These smaller bands could

be smaller Vtg mRNA transcripts, specific mRNA degradation products, or alternative

splice variants. It is possible that they represent highly similar mRNAs that contain

sequences that are homologous to our probes. Belonging to the apolipoprotein gene








family (Wahli et al. 1981), Vtg may share significant homology in some regions with

other proteins in the family. This is not likely since the bands persisted following very

stringent washing at 68C. Other studies report similar smaller Vtg transcripts. At least 4

small putative Vtg transcripts at 3.8, 3.1, 1.9, and 1.3 kb have been reported in several

fishes from the family Cichlidae (Lee et al. 1992).


Exposure of Adult Largemouth Bass to Estradiol

The first experiment on adult (>1 year old) LMB was conducted to establish a

time course of ER and Vtg mRNA induction to E2 relative to plasma Vtg accumulation.

Plasma estradiol levels were about 2500 pg/mL plasma at each of the time points

collected: 6, 12, 24, and 48 h post injection (Figure 2-11). Data from a single E2-

injection of fathead minnow (FHM) seemed to result in decreasing levels of plasma E2

over time, returning to baseline by 48 h (Korte et al. 2000). It is not clear why plasma E2

did not decline rapidly between 6 to 48 h following injection in LMB, but this may reflect

a slower clearance from the plasma in LMB than in fathead minnow. In addition to

species variation, different carrier solvents were used in for the two exposures, DMSO

(LMB) and ethanol/corn oil (FHM), and that may account for the clearance differences

between the two experiments. In that FHM study 70% of the injected E2 is quickly

cleared by 8 h post injection indicating rapid assimilation into the bloodstream (Korte et

al. 2000). That initial bolus of E2 (thought to have occurred in LMB as well) probably

led to the increase in ER and Vtg mRNA observed in LMB.

Over the forty-eight hours following the 2 mg/Kg injection in male adult LMB,

ER and Vtg mRNA increased to observed maximum amounts at 48 h of 3.68 and 533

pg/gg total RNA respectively. This pattern of Vtg mRNA accumulation over 48 h is

supported by studies in rainbow trout (Le Guellec et al. 1988), fathead minnow (Korte et








al. 2000), and sheepshead minnow (Bowman et al. 2000). The increase of both ER and

Vtg over 48 h is consistent with another report in rainbow trout (Pakdel et al. 1991).

Although the pattern of induction over 48 h is similar, the measured amounts of ER and

Vtg mRNA are about 10-fold lower in LMB compared to rainbow trout (Pakdel et al.

1991). According to Figure 2-12, LMB ER mRNA appears to accumulate faster than Vtg

mRNA up to 48 h. To better illustrate this, Figure 2-21 replots this raw data over the

initial 48 h. Figure 2-21C expresses the mRNA induction as a percent of the maximum

response, this demonstrates that ER mRNA increases faster at 12 h than Vtg mRNA.

This is even more obvious if the change in mRNA accumulation over time is plotted

against time alone (Figure 2-21E). These data are consistent with a primary response for

ER mRNA induction followed by a delayed primary response of Vtg mRNA (Figure 2-

1) over 24 h as demonstrated previously in rainbow trout (Pakdel et al. 1991).

The accumulation of plasma Vtg was also measured both qualitatively and

quantitatively. Western blots using specific monoclonal antibodies show the increase of

two bands in exposed samples (Figure 2-13A). It is not clear at this time if the two bands

represent distinct Vtg proteins or modifications of the same protein. However, one of our

two LMB monoclonal antibodies only detects one band, consistent with two distinct

amino acid sequences between the two bands. The use of monoclonal antibodies and

species-specific purified protein in the ELISA assay is more sensitive and specific than

using polyclonal antibodies with general standards (as discussed by Korte et al. 2000

using fathead minnow). This is because it measures a unique amino acid signature

(epitope) on a protein resulting in a one antibody per protein quantification. By both

Western and ELISA, plasma Vtg appeared to mimic the increase in Vtg mRNA over

time, without much delay in time (Figure 2-13). This was unexpected since other studies








have shown the accumulation of plasma Vtg to be delayed relative to Vtg mRNA

(Bowman et al. 2000, Korte et al. 2000).

By characterizing the in vivo increase in both Vtg mRNA and protein, it is

possible to monitor different levels of the ER-mediated pathway. Together these

represent activation of transcription in the nucleus to mature protein secreted into the

plasma of the animal. It has been mentioned previously that Vtg mRNA and protein

serve as distinct biomarkers of estrogen exposure in male oviparous vertebrates over time

(Bowman et al. 2000, Korte et al. 2000).


Plasma Steroid and Vitellogenin Dose Response to Estrogens

The purpose of this experiment was to test the sensitivity of adult LMB to various

estrogens 48 h following acute exposure. Specifically, a full dose response was tested for

estradiol, while smaller ranges of dose were examined for ethinylestradiol (EE2),

nonylphenol (NP), methoxychlor (MXC), and op -dichlorodiphenyl-trichloroethane

(DDT). The apparent impact of these chemicals on plasma sex steroids 48 h post

injection was minimal with the following exceptions. Forty-eight hours following the 5

mg E2/Kg dose, plasma 11 -KT approximately doubled and plasma E2 increased

approximately 4-fold (Figure 2-17), both effects were significant over controls (P < 0.05).

This effect of the highest dose of E2 (5 mg/Kg) is interesting, as it suggests that lower

levels of injected E2 are quickly metabolized and cleared from the fish in order to

maintain E2 homeostasis. The increased levels of plasma E2 at the lower doses (0.0005

and 0.005 mg/Kg) but not the intermediate doses (0.05 and 0.5 mg/Kg) at 48 h is difficult

to explain, but could be an artifact of low sample numbers (n=5) per group. These higher

levels of E2 may overwhelm the system resulting in a change in the steady state levels of

the hormone, which could have profound biological effects. The increased levels of 11-








KT, most noticeable at the highest dose of E2 (Figure 2-17B) is also interesting since the

previous experiment showed no change in plasma testosterone at 48 h following 2 mg

E2/Kg-injection. The reason for this difference in testosterone and 11-KT levels 48 h

post E2-injection is not clear at this time. This effect of E2 on 11 -KT does suggest that

exogenous estrogen exposure is capable of affecting the levels of other plasma sex

steroids, perhaps through feedback control of steroid biosynthesis.

None of the xenoestrogens tested (EE2, NP, MXC, or DDT) affected plasma 11-

KT levels at 48 h post injection in LMB. It was interesting that plasma E2 was increased

by NP in a dose dependent manner up to 600 pg/mL (Figure 2-17A), with both the higher

doses significantly higher than controls (P < 0.05). This phenomenon is not easily

explained, since NP is a weak ER agonist (Petit et al. 1997) acting much farther

downstream than E2 synthesis or metabolism. These data support the possibility that the

biological activity of NP is not just at the level of the ER, but may also affect

steroidogenesis or steroid metabolism. This increase in plasma E2 following NP

exposure is consistent with a similar observation in fathead minnow (Giesy et al. 2000),

but conflicting with a more recent study by the same lab using nonylphenol ethoxylates

(Nichols et al. 2001). This would indicate that this effect on plasma E2 is unique to the

final breakdown product (NP) of nonylphenol polyethoxylates. Another study using

flounder exposed to octylphenol (2 mg/Kg), another alkylphenol, increased plasma E2 as

well (Mills et al. 2001). However, a study in Atlantic salmon reported that 5 mg/Kg NP

actually decreased plasma E2 2 weeks following injection (Arukwe et al. 1997). These

data would indicate that this apparent increase in plasma E2 by NP may be dependent on

the species and the time of measurement.








All doses of MXC appeared to increase plasma E2 compared to controls (P <

0.05) by 48 h post injection (Figure 2-17A). This increase in plasma E2 by MXC was

also unexpected, but it was not dose dependent at the doses and time selected. Therefore

it seems to be impacting steroidogenesis or steroid metabolism in a manner unique from

NP. There is evidence that NP and MXC induces cytochrome P450s, specifically

CYP3A (Lee et al. 1996a, Li et al. 1995). NP even seems to decrease CYPlA mRNA

levels and inhibit 7-ethoxyresorufin 0-deethylase activity (Lee et al. 1996b). How MXC

appears to induce CYP2B and 3A has been investigated (Li & Kupfer 1998), but little has

been done to determine the physiological effects of this metabolic activation. There is

also evidence that MXC inhibits the enzymatic activity of CYP3A, probably inactivating

the induced enzyme (Li et al. 1993). There is also evidence that MXC inhibits, but not

inactivates, CYP2AI, 2BI/B2, and 2C I1 (Li et al. 1993). Specifically the inhibition of 2-

hydroxylation of E2 by MXC (Li et al. 1993) may result in the increased levels of plasma

E2 in this study. More studies into this possible mechanism are obviously warranted,

especially since NP and MXC appear to be affecting plasma E2 levels differently. Either

way, the accumulation of plasma E2 is probably a function of altered metabolism by these

xenoestrogens. Plasma estradiol levels were not affected by any doses of EE2 or DDT,

even though DDT is known to induce CYP3A as well (Li et al. 1995). Doses of 30 to

120 mg/Kg DDT have been shown to not effect plasma E2 at 4 to 8 weeks following

exposure in flounder (Mills et al. 2001).

As illustrated in Figure 2-1 8A, E2-induced plasma Vtg in a dose dependent

manner up to approximately 9 mg/mL by 48 h post injection of adult LMB. The

accumulation of Vtg obtained at 48 h post injection for the 0.5 and 5 mg E2/Kg doses is

comparable to the levels obtained from the 2 mg/Kg dose tested previously in adult LMB








(4.5 mg/mL). The 0.05 mg/Kg was the lowest dose of E2 that led to Vtg accumulation by

48 h post injection. The only other chemical that significantly induced plasma Vtg by 48

h was EE2 (Figure 2-18B). Surprisingly the lowest dose of EE2 (0.005 mg/Kg) induced

plasma Vtg accumulation up to 4 mg/mL. This data demonstrates that EE2 is much more

potent than estradiol in LMB, which is consistent with Vtg induction (Nimrod & Benson

1996) and binding data (Nimrod & Benson 1997) reported in catfish. The higher doses of

EE2 resulted in a dose-dependent increase in plasma Vtg.

As shown in Table 2-1, none of the doses of NP, MXC, or DDT induced plasma

Vtg over the limit of detection by 48 h post injection. Despite the relatively low doses

evaluated (0.05 to 5 mg/Kg), the lack of plasma Vtg was unexpected since these are

suspected xenoestrogens that have been shown to induce Vtg in other systems (Arukwe et

al. 1998, Hemmer et al. 2001, Nimrod & Benson 1996, Thorpe et al. 2000).

Interestingly, one study in flounder using 30 to 120 mg/Kg DDT did not show an increase

in plasma Vtg, albeit at 4 to 8 weeks following exposure (Mills et al. 2001). One

explanation is that these weak xenoestrogens may be inducing synthesis of Vtg, but it

would not be detectable until after 48 h. A more likely explanation however may be

related to the dose tested. In the previous studies the doses tested ranged from 25 to 380

mg/Kg for NP, MXC, and DDT (Arukwe et al. 1998, Nimrod & Benson 1996, Yadetie et

al. 1999). In particular, dose response by NP in salmon demonstrated that 5 mg/Kg did

not induce plasma Vtg by 1 or 4 days post injection, and 25 mg/Kg did not induce Vtg

until 4 days (Yadetie et al. 1999). Therefore, the LMB data are consistent with the

potency of these xenoestrogens in vivo as presented by other investigators using this route

of exposure. According to ER binding studies in catfish, all three of these three

xenoestrogens were 1000 times less potent than E2 (Nimrod & Benson 1997). Therefore








based on the E2 data in LMB, a dose of> 50 mg/Kg would be needed to induce plasma

Vtg. These dose response studies demonstrated that plasma E2 levels were more sensitive

than plasma Vtg when challenged by xenoestrogens at acute doses up to 5 mg/Kg.


Exposure of Juvenile Largemouth Bass to Estradiol

One acute estradiol-injection study in rainbow trout reported that peak levels of

Vtg mRNA actually occurred a week following the peak levels of ER mRNA (Pakdel et

al. 1991). Since the first time course tested in LMB was only out to 2 days, a longer time

course was conducted over 21 days in order to find where the peak expression levels

were for these genes in LMB. This study was completed using juvenile LMB (< 1 year

old) undergoing their first reproductive cycle as illustrated by histology of their gonads

(Figure 2-4C and 2-4D). This gender distinction made it possible to report data for male

LMB only.

Northern and slot blot analysis was used to characterize the gene expression

pattern of ER and Vtg mRNA following the single E2-injection (2 mg/mL). Both ER and

Vtg mRNAs accumulated to peak levels by 48 h post injection (Figure 2-19). At four

days ER mRNA levels had reached a plateau, whereas Vtg mRNA levels were still

returning from peak levels. This time point, perhaps more than any other, best

distinguishes the temporally distinct pattern in mRNA levels between ER and Vtg

mRNA. This difference between the levels of mRNAs is probably a combination of

unique mRNA half-lives in addition to the changing rates of synthesis and degradation.

By Day 7 through Day 21, both mRNAs had returned to levels initially observed at 6 h

post injection. As illustrated in Figure 2-19 the peak amount of ER mRNA produced (3.5

pg/ltg total RNA) is only a fraction of the amount of Vtg mRNA (137.8 pg/gg total

RNA) induced by the same treatment. Neither mRNA had returned to basal levels by








three weeks post injection. The estradiol levels remaining after 6 h (as shown in Figure

2-11) could be responsible for the maintenance of above normal, but low levels of these

mRNAs out to 21 days. Although plasma estradiol was not measured in this particular

experiment, previous data from fathead minnow (Korte et al. 2000) would suggest that

the initial bolus of E2 following the injection caused the initial increase in ER and Vtg

mRNA observed, but that the lower level of E2 remaining out to 48 h (Figure 2-11) and

perhaps longer could help explain why neither mRNA returned to basal levels by 21 days

in LMB.

As shown in Figure 2-19, both ER and Vtg mRNA peaked at 2 days post

injection. These data are similar with the previous experiment in adult LMB. This peak

of Vtg mRNA induction at 2 days and declining over time is supported by studies using 2

to 5 mg E2/Kg in rainbow trout (Le Guellec et al. 1988), fathead minnow (Korte et al.

2000), and sheepshead minnow (Bowman et al. 2000). The observation in rainbow trout

following single E2-injection (Pakdel et al. 1991) that reported ER mRNA peak levels

(Days 2 to 6) occurring a week prior to Vtg mRNA peak levels (Day 15) is not consistent

with these results in LMB or with other Vtg rnRNA induction studies cited earlier. It is

not clear why the Pakdel study observed such late peak levels of ER or Vtg mRNA.

There are many factors that may be responsible for the apparent discrepancy in results

between the LMB data and that presented by Pakdel. One possibility is the carrier

vehicle used to administer the E2, Pakdel reports using saline compared to cocoa butter

(rainbow trout) (Le Guellec et al. 1988), ethanol/corn oil (fathead minnow) (Korte et al.

2000), triethylene glycol (sheepshead minnow) (Bowman et al. 2000), or in this

experiment, dimethylsulfoxide (LMB). The choice of carrier may dramatically affect the

uptake, disposition, and rate of release of the chemical over time, especially comparing








saline versus these other carriers. Other factors to consider are the temperature and salt

maintenance differences between these freshwater and marine species (rainbow trout at

160C compared to 25C for LMB). The reproductive status of the fish is also an

important consideration when interpreting and comparing hormone-dependent results

between experiments as discussed later, but certainly when evaluating data across

species.

The primary hypothesis for these LMB E2-injection studies was that estradiol

would induce a primary and delayed-primary hepatic transcriptional response over time

for estrogen receptor and vitellogenin mRNAs respectively. Initially this was based on

the observation of temporally distinct (days) peak levels of ER and Vtg mRNA reported

in rainbow trout (Pakdel et al. 1991). On closer examination of how a primary and

delayed primary response gene was defined, the temporal distinction between the two

responses should only be hours, not days (Dean & Sanders 1996). By mechanistic

definition, ER and Vtg transcription represent primary and delayed primary responses.

What this means is that ER synthesized by the primary response is partially responsible

for the transactivation of the Vtg gene (hence a delayed-primary response). Both

estradiol-injection time course experiments were designed to test this idea that ER and

Vtg mRNA fall into these distinct response categories. As described in detail below, the

LMB mRNA data supports the hypothesis that the transcription of ER and Vtg represent

primary and delayed primary responses respectively.

The most important time frame for distinguishing these responses are their

respective rates of mRNA accumulation compared to the time in which they reach their

peak amounts. For both adult and juvenile LMB the ER and Vtg mRNA responses peak

at 2 days post injection. Figures 2-21A and 2-21B illustrate the raw data for the








induction of these mRNAs up to their peak levels observed at 2 days. Because both

studies were conducted using approximately the same dose, Figure 2-21 also illustrates

how age (juvenile vs. adult) can affect the amount of mRNA synthesized. In adult LMB,

there appears to be a much lower basal level of ER mRNA, but a higher dynamic range of

induction compared to juvenile fish. Conversely adult LMB seem to have higher basal

levels and induction capacity of Vtg mRNA than juvenile fish. In addition to age, the

differences seen between the two experiments can also be attributed to the reproductive

status of the fish. The juvenile fish appeared reproductively active in their first season,

but were sampled in January (moderate to high spermatogenic activity) (Figure 2-4C)

compared to the adults which were sampled in July (low spermatogenic activity) (Figure

2-4A). It is not known at this time if reproductive status or age may impact the current

hepatic mRNA results in LMB, but this is currently being investigated in the lab.

The magnitude of Vtg mRNA induction being so much greater than that of ER

mRNA has also been observed previously in rainbow trout (Flouriot et al. 1996, Flouriot

et al. 1997, Pakdel et al. 1989, Pakdel et al. 1991). The reason for the different

sensitivities and mRNA response to the same E2-exposure may be the ability of the

activated ER complex to bind the gene-specific estrogen response elements (ERE). The

ERE in the promoter of the ER gene is thought to be imperfect (perhaps just a half-site),

which may help explain the weak induction of this gene compared to the perfect ERE

thought to exist in the promoter of the Vtg gene. In addition, the rainbow trout ER gene

has a low threshold response to ligand-bound receptor and increasing amounts of ER

protein do not affect ER mRNA following E2-exposure (Flouriot et al. 1997). This is

consistent with it being a primary mRNA response.








Post-transcriptional stabilization of the induced mRNAs are also thought to

contribute to the levels of ER and Vtg mRNA present in the cell (Flouriot et al. 1996,

Shapiro et al. 1989). As reported in vitro for rainbow trout using actinomycin D, the

half-life of ER mRNA is 4 h and Vtg mRNA is 10 h, however in the presence of E2 these

half-lives were extended to 14 and 29 h respectively (Flouriot et al. 1996). This effect of

estrogen on the post-transcriptional stabilization of ER and Vtg mRNA is consistent with

other reports in human breast cancer cells (Saceda et al. 1998) and frog (Brock & Shapiro

1983, Dodson & Shapiro 1997). Based on results in LMB, the low level of E2 remaining

in the plasma following injection could be responsible (through mRNA stabilization) for

these mRNAs not returning to basal levels by 21 days post injection (Figure 2-19). The

end result is that these genes have very specific, but different sensitivities to the

concentration of E2 and ER protein present in the cell at any particular time (Flouriot et

al. 1996, Flouriot et al. 1997).

Due to this large difference in the magnitude of the mRNA accumulation between

the two genes ER and Vtg mRNA it is difficult to illustrate the distinction between the

hypothesized primary and delayed primary responses. Therefore the raw data (Figure 2-

21A and 2-21B) is replotted as a percent of the maximum response at 2 days for each

mRNA per experiment (Figure 2-21C and 2-21D). This replot illustrates how LMB ER

mRNA appears to be up regulated as a higher percent of its peak level earlier than Vtg

mRNA (ER mRNA curve is slightly to the left of Vtg mRNA). Perhaps a more

appropriate comparison between the accumulations of these genes is the rate of mRNA

accumulation using derivative of the raw data (change in mRNA over time/time) as

shown in Figure 2-21E and 2-21F. This replot suggests the rate of ER mRNA peaks

much earlier (6 to 12 h) than Vtg mRNA (48 h) following acute E2-exposure.








The results of the analysis in Figure 2-21 is consistent with reports with rainbow

trout E2-exposure where the transcription rate of ER mRNA peaks at 6 to 12 h and peak

rates of Vtg mRNA follow at 48 h (Flouriot et al, 1997). Specifically it appears that the

ER gene has a high sensitivity, low capacity for E2-induction, whereas Vtg has a low

sensitivity, high capacity for E2 (Flouriot et al, 1997). Since ER protein accumulation

was not measured in LMB during this time, it is difficult to prove that the ER mRNA

induced earlier is partially responsible for the short time lag in Vtg mRNA induction.

But these LMB data do still support the hypothesis that ER and Vtg mRNA are primary

and delayed primary response genes. In fact, an earlier study in rainbow trout

demonstrated that ER mRNA does actually appear at least 6 h earlier than Vtg mRNA

following single E2-injection (0.5 mg/Kg) (Pakdel et al. 1989). As described earlier,

another study in rainbow trout also supported this hypothesis, albeit on a much longer

time scale, and using actual peak times for each mRNA (Pakdel et al. 1991). All these

data support a very tightly regulated interaction between these different genes dependent

on age, gender, time, and concentration of hormone. Future studies in this field should

identify how estrogen mimics can interfere with endogenous primary and delayed

primary responses, and how those disruptions can manifest themselves in vivo, probably

during reproduction and development when these responses are thought to be most

sensitive.


Differential mRNA Regulation by Estradiol

Estradiol is thought to regulate many physiological actions, most of which are

thought to occur through the ER. Because the ER itself is auto-regulated, this serves as

an additional level of interaction between E2-regulated genes. The primary mechanism of

E2 action is through the ER, which then binds to estrogen response elements leading to up








or down regulation of a variety of genes. Unfortunately only a handful of genes have

actually been identified that are regulated by E2-ER directly. To this end, differential

mRNA display RTPCR was done to identify the fingerprint of E2 gene regulation over

time, as well as to possibly identify novel targets of E2-regulation.

As shown in Figures 2-14 and 2-15, there are many changes in mRNA expression

that occur over 48 h. Arrows with double lines on the right indicate mRNA populations

that are transiently up regulated and regular arrows on the left point out mRNAs down

regulated over 48 h (Figure 2-14). In Figure 2-15, regular arrows on the right indicate up

regulated mRNAs and stippled arrows on the left represent mRNAs not affected by E2-

exposure over 48 h. Several of the bands of interest (indicated by arrows) were excised

from the gel, cloned, and sequenced. So far only two mRNAs have been positively

identified. The first was Vtg, and as described earlier this was validated extensively as

estrogen inducible. The second mRNA identified, a protein disulfide isomerase-related

protein termed ERp72, appears to be a novel target of estrogen transcriptional regulation.

The predicted amino acid sequence of LMB ERp72 aligns with high similarity to

the human, rat, and mouse ERp72 sequences (Figure 2-16A). Because this sequence is at

the 3'-end it is possible to identify the retention signal at the carboxy-terminus of the

protein (-KDEL) that is diagnostic of this family of endoplasmic reticulum membrane

bound proteins (Mazzarella et al. 1990). This family of proteins were initially

characterized by their redox capacity in catalyzing disulfide bond formation (Freedman,

1984). There have been many other functions attributed to this family of enzymes

including thioredoxin activity (binding to thyroxine) (Freedman, 1989), binding to a

variety of sequence non-specific peptides (Noiva et al. 1993, Spee et al. 1999); binding to








calcium (Van et al. 1993), assembly and folding of antibodies (Iida et al. 1996) and

presumably binding to the estrogen receptor itself (Landel et al. 1994).

Another function of ERp72 is assisting in the folding and lipidation of

apolipoprotein B (Linnik & Herscovitz 1998). This is very interesting since Vtg belongs

to the apolipoprotein gene family (Wahli et al. 1981). If ERp72 was partially responsible

for folding Vtg, then it would be expected to be up regulated by E2 to accompany the

increased synthesis of Vtg. In fact, it was reported that this family of protein disulfide

isomerase proteins are preferentially expressed in cells actively secreting proteins

(Hayano & Kikuchi 1995). Even more interesting, but perhaps confusing is a report

providing evidence that the ERp72 actually binds estradiol. In fact a portion of the

ERp72 amino acid sequence actually shares two spans of high homology to the hormone

binding domain of estrogen receptor (Tsibris et al. 1989).

This LMB ERp72 cDNA clone was used as a probe in Northern blot experiments

to validate its estrogen regulation (Figure 2-16B). It appears that LMB ERp72 is very

rapidly induced following E2-injection. In fact it seems that it is maximally induced by 1

day post injection, and quickly returns to just above basal levels by 2 to 4 days (Figure 2-

16B and 2-16D). This is intriguing given the context of what was investigated regarding

ER and Vtg mRNA induction in these same samples. It would appear that ERp72 may

not require the induction of ER mRNA to synthesize its peak levels like Vtg mRNA is

thought to. If ERp72 serves some function in Vtg protein folding, then it makes sense

that it is up regulated prior to Vtg mRNA synthesis. Very limited information is known

on how ERp72 is regulated (Dorner et al. 1990). There is some evidence that another

protein disulfide isomerase family member, ERp61 (HIP-70), is under hormonal control

(Kaplitt et al, 1993). Since ERp72 has not been previously documented as estrogen-









inducible or even identified in teleosts, much more characterization needs to be done to

better understand its transcriptional regulation and function in the overall estrogen

response. In addition to ERp72, the other E2-regulated mRNAs illustrated in Figures 2-

14 and 2-15 need to be identified and characterized, perhaps leading to a better

understanding of the family of responses to estradiol.


Table 2-1. Plasma Vtg levels 48 hours post E2 injection

dose* E2 EE2 NP MXC DDT
0.00 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000)
0.0005 0.004 (0.009) N/A N/A N/A N/A
0.0050 0.018 (0.025) 4.358 (1.602) N/A N/A N/A
0.0500 0.480 (0.402) 7.978 (1.743) 0.005 (0.005) 0.000 (0.000) 0.007 (0.010)
0.5000 3.310 (1.471) 11.783 (2.043) 0.002 (0.006) 0.003 (0.008) 0.007 (0.022)
5.0000 8.846 (1.035) N/A 0.000 (0.000) 0.000 (0.000) 0.003 (0.006)

*Dose is mg chemical/Kg body weight (ppm)
Data are presented as the mean nig vitellogenin/mL plasma (standard deviation)
N/A is not applicable




























Largemouth bass mRNA coordination hypothesis.
Following acute estrogen stimulation, ER and Vtg
mRNA would be represent primary and delay
primary responses, respectively, over time.


Figure 2-2. Largemouth bass (Micropterus salmoides).


primary response

_______________ ER mRNA 4T


0 + 0 ER protein


_pVtmNA


delayed-primary response


Figure 2-1.



























00










Figure 2-3. Largemouth bass sampling by lab personnel. A) 150 gallon tanks used for experiment; B)
Largemouth bass in a tank; C) Kevin Kroll collecting first sample; D) Assembly line used
to process samples on site.













































Histology of largemouth bass gonads. A) Male LMB
sampled in July with low to no observable spermatogenic
activity; B) Male LMB sampled in November with low to
moderate spermatogenic activity; C) Male LMB sampled
in January with moderate to high spermatogenic activity;
D) Female pre-vitellogenic oocytes.


Figure 2-4.













ci


c i


I + -I


Estrogen receptor and vitellogenin RT-PCR products.
RNA is from in vivo E2-exposed largemouth bass
samples, all of the expected size. A) RT-PCR using ER
specific primers on E2 and control RNA; B) RT-PCR
with Vtg primers on control, 24 and 48 h post E2-
injection. Size markers on the left of both gels are as
follows from top to bottom: 1000, 750, 500, 300, and
150 bp.


Figure 2-5.


E2I



















*clone RT-PCR product into
multiple cloning site


sequence If pUEM
using M13 -3k
for/rev
primers

Digest for cDNA

probe

IECoRI I VTG ECoRI


PCR offplasmid for in vitro
\transcription template (std.)


M13F T7 sense VTG I M13R]


Cloning strategy for estrogen receptor and vitellogenin
PCR products. The system was used to generate
templates for cDNA probes, cycle sequencing with M13
primers, and in vitro transcription of cRNA standards.


Figure 2-6.














A B C




2e+07
r=0.99

A"POW le+07


0



-, ->N 1000000 ....."''"
0.050.1 1 10 100 1C
ng Vtg cRNA



Figure 2-7. Largemouth bass vitellogenin cRNA standard curve.
Performed using in vitro transcription; A) Methylene
blue stain of in vitro synthesized cRNA standards
(serially diluted) prior to hybridization; B) Standards
probed with VTG cDNA probe (following
hybridization); C) Data recovered from phosphorimager
plotted against the known cRNA concentrations to
create a standard curve for that specific blot, r= 0.99.



















LMB-ER
Tilapia-ERa
Seabream-ERaO
Trout-ERa
Human-ERa
Mouse-ERa
Rat-ERa

LMB-ERa
Tilapia-ERa
Seabream-ERa
Trout-ERa
Human-ERa
Mouse-ERa
Rat-ERa

LMB-ERa
Tilapia-ERa
Seabream-ERa
Trout-ERa
Human-ERa
Mouse-ERa
Rat-ERa


LMB-ERa
Tilapia-ERa
Seabream-ERa
Trout-ERa S
Human-ERa DHI
Mouse-ERa DHII
Rat-ERa DHII




Figure 2-8.


Fl
TA
SA
SA


LT Q
TLT
LI


Multiple sequence alignment of cloned largemouth
bass estrogen receptor fragment. Predicted amino
acid sequence for largemouth bass ER from PCR
amplified sequence.






















LMB-Vtg
Mummichog-Vtg
Tilapia-Vtg
Trout-Vtg
Chicken-Vtg
Frog-Vtg

LMB-Vtg
Mummichog-Vtg
Tilapia-Vtg
Trout-Vtg
Chicken-Vtg
Frog-Vtg


LMB-Vtg
Mummichog-Vtg
Tilapia-Vtg
Trout-Vtg
Chicken-Vtg I
Frog-Vtg I




Figure 2-9.


RE

KNN
EIHGVIj
EGVIm

NGQVKL
SADAFKM


E SN K S.I
*A5SEISLISAG.AL
NLISEEHLEKSFNYPTVI


ITIP
; FKFRPT


H SE Q DGS DQMRFI I PAEJI




Multiple sequence alignment of cloned largemouth
bass vitellogenin fragment. Predicted amino acid
sequence for largemouth bass Vtg from PCR
amplified sequence.










ER mRNA

C E,


VTG mRNA

C IE


kb
5.0-*

3.3--*
, 3.5
1.7_-.




2.1




3.4

1.9


Figure 2-10. Largemouth bass liver mRNA characterization. Northern blots of
control and exposed livers in duplicate. A) Estrogen receptor and
vitellogenin mRNAs; B) P-actin mRNA; C) rRNAs stained with
methylene blue.

















3000

E 2500
CO
CO
C" 2000
-J
E
0) 1500
0L

1000

C 500

0


0 6


12 24 48

hours estrogen
Testosterone


Figure 2-11.


Plasma steroid measurements out to 48 h post injection.
These fish were adult LMB injected once with 2 mg/Kg
E2. Male LMB sampled in July with low to no
observable spermatogenic activity.














I-DMSO 6 1 12 1 241 48 0


1-. 5


. 2-


0
ICL8


............. I ................... ........... 1-..... ............
....... ....................... .20 .........................


700
600-
500 -
400
300
200
100
0


0 10 20 30 40 50 60

hours post injection


Figure 2-12. Estradiol induced mRNAs over 48 hours. A) ERa mRNA
mRNA by Northern; B) Vtg mRNA by Northern; C) 13-actin
mRNA by Northern; D) ERa mRNA corrected to P3-actin by
slot blot; E) Vtg mRNA corrected to P3-actin by slot blot.


A


q*I


zC
c





-
"-0


I
















A


DMSO 6 12 24 48 0



5a


1

0.1

0.01

0.001


10 20 30


40 50


hours post injection


Figure 2-13. Plasma vitellogenin induction over 48 hours. A)
Western blot demonstrating the two Vtg monomers
(180 and 160 kDa) induced by estradiol; B)
Quantitation of Vtg induction by ELISA using the
same set of samples.


..! ... .


. . . . . . . . . . . . . . . . . . . . . . .









































4=
4=


Figure 2-14.


Largemouth bass differential display primer pair G-23.
Arrows on left represent mRNAs down regulated over time
following E2 treatment. Arrows on right represent transient
increases in specific mRNAs. Three separate fish liver RNA
(lanes) analyzed per timepoint. Samples were taken over 48 h.











0 12 24 48



L
*0*OW


B



0*





*0O


4--




4 Vtg




I= ERp72













,4-


Figure 2-15.


Largemouth bass differential display using primer pairs G-10
and C-i Arrows on left represent mRNAs constitutive over
time following E2 treatment. Arrows on right represent up
regulation of specific mRNAs. Vtg and PDI are two of the
identified mRNAs up regulated by E2. Three separate fish
liver RNA (lanes) analyzed per time point. A) Primer pair G-
10; B) Primer pair C-1. Samples taken over 48 h.












LMB-ERp72
A HUMAN-ERp72
RAT-ERp72
MOUSE-ERp72

LMB-ERp72
HUMAN-ERp72
RAT-ERp72
MOUSE-ERp72


S 10.25 1 2 1 2 14 17 1 1


I W I- 2.6kb


-4- 2.1kb



5e+5
. &. 4e+5
S3e+5-
W-
m 2e+5.
le+5.
M 0 051 9 4 7 14 91


days post injection


Figure 2-16.


Verification of largemouth bass ERP72. A) Multiple sequence
alignment of predicted amino acid sequence lined up against
some mammalian species; B) E2-induction of LMB ERp72
mRNA over 21 days by Northern analysis; C) 13-actin mRNA
on same Northern blot; D) Relative quantitation of ERp72
mRNA induction by E2.


GK

G K

)KR
U.TKE*
TKE*
kTKEJ


Vob I








1200 *
- 1 0 0 0 .............................. ....................
WJ .0.0005 mg/ kg
E 800 *. ......................* 0.005mg/kg
E 6 .g..................... ..... ..... ................... ....... ................ 0 .0 5 m g k g
i c 400 ..................... . 0.5 mglkg
-- 200 .5.0 mg/kg
0L 0
Control E2 EE2 NP MXC DDT
1200
1 0 0 0 ... ....................... ............................ ...................................
E 800 .......................... ... ........ 0.005 mg/ kg
CU-00- ....... 0.005 mg kg .
600 ............................................. ..0.05mg
E L 400 ..... ... .5 mg/kg
200 5.0 mg/ kg

Control E2 EE2 NP MXC DDT

Figure 2-17. Plasma steroids 48 hours following exposure. All data is from male adult LMB. A) Plasma estradiol; B)
Plasma 11 -ketotestosterone (11 -KT). Astericks indicate experimental groups that were significantly
different from controls (P<0.05) using the Student's t-test.





73




14

A 1 2 .................................................................................................
_J
E 10 .....................................................................


C ) 84 ............................................................... .....
E





....... 2..................................................
0

-2 1 1
0.0001 0.001 0.01 0.1 1 10
dose (mg/Kg)
14
B

12..................................... ....
-J










2,,
0.01 0.1
dose (mg/Kg)

Figure 2-18. Dose response induction of plasma vitellogenin in
largemouth bass. A) Estradiol-induced plasma
vitellogenin at 48 h post injection; B)
Etbinylestradiol-induced plasma vitellogenin.











0 10.25 1 2 1 4 1 7 1 14 1 21 1


o 5 10 15 20 25
days post injection








0 5 10 15 20 25
days post injection


Figure 2-19.


Time course of estradiol induced mRNAs over 21 days. A) ERa
mRNA mRNA by Northern; B) Vtg rnRNA by Northern; C) 0-
actin mRNA by Northern; D) ER ac mRNA corrected to 1-actin
by slot blot; E) Vtg mRNA corrected to fP-actin by slot blot.


C
Zc
"a0
Q.,c&


< t
Z M
-7


CL8































0I I I
0 5 10 15 20


days post injection


Figure 2-20.


Time course of plasma vitellogenin induction.
Quantitation of vitellogenin by ELISA.


1



0.1



0.01



0.001










Adult/summer


100



o10


E8
E


1000



0
~ 0
< 1


WZ8
E
0.1



100 .


EmRNA

o 1 2
days post injection




............... V:..................................


.. ... . ~ _R....... .... ...... ............. ......... -
..... .. ..... ..... ..... .. . ....... .......... V ig .....


60 ..


Juvenile/spring









-V V9f~NA


days post injection


0 1 2
days post injection


4)
E 40
Z 3
E 2
<


days post injection


Figure 2-21.


days post injection


Largemouth bass estrogen receptor and vitellogenin mRNA
comparison. The induction of ER and Vtg mRNA over 2
days are illustrated together three different ways. The data
on the left is from adult male LMB taken during summer.
The data on the right is from juvenile male LMB taken
during spring. A) and B) raw mRNA data; C) and D) mRNA
data plotted as a percent of the maximal response; E) and F)
mRNA data plotted as the increase in mRNA divided by
time.


0 1 2
days post injection


D


-0- ER n~NA
~


.... V l* ..............


......... .

.. ... ... .....

I ER
-11 V NA













CHAPTER 3
DEVELOPMENT OF AN IN VITRO MODEL: LARGEMOUTH BASS PRIMARY
HEPATOCYTES


Introduction

The use of fish primary hepatocytes as a model system represents a relatively new

approach to characterizing the cell biology and molecular physiology of teleosts (Baski &

Frazier 1990, Hightower & Renfro 1988, Moon et al. 1985). Part of understanding gene

conservation and evolution involves cross species biology. The significance in

understanding an environmentally relevant species (such as largemouth bass) on a

molecular level will help explain how xenobiotics may be impacting our environment.

The existing approaches to this problem include field studies, very limited in vivo

controlled exposures, and more recently, recombinant yeast expression assays. The field

studies are almost entirely observational and do not provide detailed insight into possible

mechanisms of xenobiotic disruption. Controlled in vivo exposures are limited because

of the availability of healthy tank-raised fish. Recombinant yeast assays continue to be of

limited value because of incomplete representation of tissue metabolism and intracellular

protein constituents, both present in vertebrates and critical to target gene regulation.

Fish primary hepatocytes were initially developed and characterized using

goldfish and rainbow trout (Baski & Frazier 1990). More recently, the demand for new

models has led to cultures being isolated from various other aquatic organisms such as

eel, catfish, and flounder. In a more general sense, largemouth bass may serve as a

sentinel species in field population data because of its predatory nature and apparent








environmental sensitivity. However, studies using largemouth bass (LMB) as a model

are very scarce in the literature, despite its importance in sports and recreation in the

southeastern United States. Following development of this assay, it is anticipated that

xenobiotic studies can be conducted using LMB and will constitute the first report on the

characterization of this fish at the molecular level.


Rationale

One of the primary mechanisms of action for hormones is to bind to cognate

intracellular receptors and induce de novo transcription. Based on this, most of the in

vitro assays used to assess estrogenicity (the most studied mechanism of endocrine

disruption) rely on this ligand-receptor mediated response. These methods include

receptor binding, cell proliferation, receptor-dependent gene expression, and recombinant

receptor-reporter gene constructs (Jensen & Jacobson 1960, Jobling & Sumpter 1993,

Soto et al. 1995, White et al. 1994). Each of these assays has distinct advantages and

disadvantages (Zacharewski, 1997, Zacharewski, 1998). All these assays are useful in

determining very specific aspects of the predicted mechanism of action (receptor

binding), even in so far as to allow receptor-DNA interaction followed by transcription

initiation (gene expression and gene constructs). They are easily adapted to large scale

screening and do not require the use of animals. Cell proliferation is a simple functional

assay of estrogenicity, unfortunately there are few estrogen-responsive cell lines. While

the in vitro assays are useful, they each have disadvantages as well. Receptor binding

and a number of reporter gene constructs cannot distinguish between agonists and

antagonists. Xenobiotics that require metabolic bioactivation (proestrogens) will not be

detected by a majority of these assays, with the exception of assays using primary








hepatocytes. Most of the yeast or cell line assays cannot account for in vivo

pharmacokinetics.

The ability of in vitro assays to accurately predict in vivo responses of hormone

mimics is limited. Disruption of the endocrine system can include multiple pathways

involving different organs with unique chemical sensitivities at critical windows in time.

Because of the complexity in potential pathways, it is difficult to isolate tissue-specific

mechanisms of action. For xenoestrogens in particular there also seems to be non-

receptor-mediated effects on the cell. As discussed earlier, ligand-independent receptor

activation can also elicit responses. The in vivo mechanisms of specific gene regulation

and expression depend on the presence and state of intracellular constituents, which are

not always accurately represented by transfected reporter assays in transformed cell lines

or yeast. In fact, using yeast or cell lines, one must transfect and express estrogen

receptors since these systems lack these most basic elements in the pathway. This begs

the question of what other regulatory cofactors are missing in these systems? Therefore,

the importance of an unaltered vertebrate system with endogenous markers is critical to

understanding how the mechanism of xenoestrogen disruption may interfere with the

reproductive and developmental success of an individual. As discussed below, primary

fish hepatocytes may fulfill this need in vitro without additional modifications and

including endogenous markers necessary to investigate specific mechanisms of

xenoestrogen action.


Fish Primary Hepatocytes

To better understand how contaminants are impacting the biological systems, very

specific molecular responses need to be quantified using an unaltered cellular system that








accurately represents in vivo end points. Fish primary hepatocytes, which retain an active

estrogen receptor and intact metabolic pathways can provide this at the cellular level

(Baski & Frazier 1990, Flouriot et al. 1993, Hightower & Renfro 1988, Moon et al.

1985). This system can also be used to measure endogenously expressed markers, such

as vitellogenin and estrogen receptor mRNAs that can be directly correlated with in vivo

experiments (Anderson et al. 1996).

To best study the mechanisms of disruption of specific xenoestrogens, an ideal

model would share the in vivo effects of bioactivation and metabolism with the feasibility

for extensive experimental manipulation provided by an in vitro system. Because

primary hepatoctyes still retain many of the functions of liver tissue, they respond

similarly. Among the functions they retain is the ability to metabolize compounds, bind

hormone active agents through the endogenous estrogen receptors, and recruit

intracellular proteins to initiate de novo transcription (Baski & Frazier 1990). This is

particularly important for contaminants such as methoxychlor that have been shown to

confound mammalian cell transactivation assays, as it requires metabolic activation

(Bulger et al. 1978, Shelby et al. 1996). Therefore the capacity of primary hepatocyte

cultures to retain more of the original organ functions than immortalized or transformed

cell lines will allow for more relevant mechanistic studies of hormone-regulated gene

expression and metabolism. The reason is that these cell lines often lose normal cellular

controls on gene regulation. Fish hepatocytes are also valuable for their importance in

steroid metabolism, as well as being the site of vitellogenin biosynthesis (Baski & Frazier

1990). This model has proven itself to be valuable in understanding various effects of

teleost liver metabolism including hormonal regulation, and protein








metabolism/expression (Moon et al. 1985). This is especially important for these studies

since fish primary hepatocytes have been shown to be estrogen-responsive and still

express an active ER, whereas cell lines do not generally share these characteristics

(Pelissero et al. 1993).

For various reasons, primary fish hepatocytes better represent in vivo cellular

conditions in fish than mammalian hepatocytes do for mammalian systems. The structure

of fish livers is not organized into a hepatic triad, rather it is remarkably homogenous.

Fish livers are also roughly 80% hepatocytes, i.e., predominantly one cell type; whereas

mammalian livers consist of only 60% hepatocytes with other cell types making up the

other 40% of the population (Blair et al. 1990, Hampton et al. 1985). Therefore fish

primary hepatocytes can potentially represent in vivo cellular and intercellular actions

very well.


Objective

The majority of information on fish primary hepatocytes has been collected using

the rainbow trout as a model. There seems to be as many different ways to isolate and

culture this cell type as there are publications on the method. Many different techniques

of isolation and culturing of hepatocytes were attempted with largemouth bass. The first

technique was using an adaptation of Ostrander and Blair from trout (Blair et al. 1990,

Ostrander et al. 1995). Because LMB hepatocytes did not thrive under those conditions,

different perfusion buffers and incubation media were used. Non-perfusion methods of

cell dissociation relying on mechanical and chemical techniques (Freshney, 2000) were

also tried. Ultimately the most promising technique was using a method developed by

Mommsen (Mommsen et al. 1994). The objective was to develop and validate a primary








hepatocyte model that could be used to examine the differential molecular mechanisms

behind gene regulation by various xenoestrogens.




Materials and Methods


Fish Collection and Maintenance

Largemouth bass (Micropterus salmoides) were purchased from American

Sportfish Hatchery (Montgomery, Alabama). They were either maintained at the Aquatic

Toxicology Facility Lab at the University of Florida under the direction of Dr. Evan

Gallagher or at the United States Geological Survey- Carribean Science Center under the

direction of Dr. Timothy Gross. Fish were held in aerated 104 to 250 gallon fiberglass

tanks, under constant conditions of 21 +/-2* C prior to hepatocyte isolation. Dissolved

oxygen, total ammonia content, and pH were monitored during this time. The fish were

exposed to ambient light concentrations, and fed Aquamax 5D05 fish feed (Purina).


Reagents and Equipment

If not specifically indicated, all reagents were purchased from Sigma-Aldrich or

Life Technologies. For hepatocyte isolation, specialized surgical tools from Roboz

Surgical (Rockville, MD) and Fisher Scientific were used. For perfusing the liver, IV-

cannulae, IV extension tubing, and silk sutures (3.0) were obtained from the nearby

hospital supply center. Collagenase Type I and IV were utilized from Sigma and

Worthington Biochemical.

For primary hepatocyte culturing, 6, 12, or 24-well Becton-Dickenson-Falcon

Primaria (Bedford, MA) tissue culture plates were used. Charcoal-stripped fetal bovine








serum was obtained from HyClone laboratories (Logan, UT). Clinical kits to determine

lactate dehydrogenase activity were used from Sigma and Promega.

RNA isolation and cDNA amplification were performed as described in Chapter 2

of this dissertation and previously (Bowman & Denslow 1999). Briefly, the individual

tissue culture plate wells were prepared for RNA using the acid phenol guanidinium-

isothiocyanate method (Chomczynski & Sacchi 1987) or using RNeasy columns from

Qiagen following the manufacturer's instructions. Total RNA isolates, resuspended in

water, were measured at 260 and 280 nm using a spectrophotometer. The 260 nm

reading was used to estimate the concentration of total RNA recovered from the isolation.

The 260/280 ratio as well as a 1% agarose-formaldehyde gel stained with ethidium

bromide were used to verify the quality of the RNA in each sample. Oligo-dT primers,

dNTPs, 5X transcription buffer (Life Technologies), and Superscript II (Life

Technologies) were used to reverse transcribe 2 pg total RNA. Vitellogenin (Vtg)

primers (80 pmol/pL) or beta actin primers (10 pmol/pL) inlOX reaction buffer with

MgC12 (Perkin-Elmer), dNTPs, AmpliTaq (Perkin-Elmer,), and 2 pl of cDNA from the

reverse transcription reaction were used to amplify portions of these respective genes by

the polymerase chain reaction (PCR). The conditions for the PCR were: hold at 801C for

3 min; hold at 94C for 3 min; 35 cycles of 94C for 45 sec, 52C for 90 sec, 72C for 45

sec; hold at 72C for 10 min; and hold at 4*C for one hour. The PCR products were

analyzed and purified by 1.2% agarose gel electrophoresis. These RT-PCR reactions

using vitellogenin or beta actin primers were performed as described in Chapter 2 of this

dissertation and in the literature (Bowman & Denslow 1999).








For both isolation and culturing of primary cells there was a necessity of several

pieces of equipment. For the perfusion of the organ, a peristaltic pump with manual

speed controls was necessary to adjust flow. For washing of isolated cells, a swinging

bucket centrifuge was necessary. Final culture and maintenance of hepatocytes required

a laminar flow biological hood, water-jacket cooled incubator, and external chiller to

maintain temperatures below ambient in the incubator. For characterization, viability,

and counting, an inverted and regular light microscope outfitted with Pixera Viewfinder

Pro (Los Gatos, CA) and analyzed using Pixera Studio Pro (Los Gatos, CA) were used.

Transmission electron microscopy was performed using a Hitachi H-7000 at 75

kV in the Electron Microscopy Core at the University of Florida. First, isolated

hepatocytes were fixed by 2% glutaraldehyde in PBS. After washing with PBS, the cells

were post-fixed in 1% OsO4 for one hour. The cells were then washed again in PBS and

water, and then dehydrated with graded ethanol, followed by acetone. The sample was

then infiltrated and polymerized by graded epoxy resin (Embed 812), followed by

sectioning with an ultra microtome. The section was then shaped using a diamond knife

to -70nm/slice. Once on a mesh, the sample was stained with uranyl acetate and lead

citrate prior to examination by the electron microscope. Sample preparation and digital

image capturing was done in collaboration with individuals in the Electron Microscopy

Core.


Buffers and Media

There are three buffered solutions necessary for hepatocyte perfusion and

isolation. All are modifications of Hanks balanced salt solutions (Hanks and Wallace,

1949). The first solution serves as a pre-perfusion buffer to remove all the blood from








the liver prior to enzymatic digestion. The second solution contains 105 to 131 U/mL

collagenase type I or IV. This bacterial collagenase (EC 3.4.24.3) is from Clostridium

histolyticum. This is used during perfusion to dissociate cell to cell interactions within

the liver. The third solution is used to wash the initial isolated cell pellet. It contains

calcium and bovine serum albumin to maintain cellular integrity during washing. All

buffers require sterile 0.2 micron filtration prior to use. See Table 1 for individual

components of the three buffered solutions.

After experimenting with other prepared media such as Dulbucco's Minimal

Essential Medium + Ham's F-12, largemouth bass hepatocytes seemed to respond best to

a modified Leibovitz's (L-15) media. Important are the bicarbonate and HEPES

buffering components to maintain necessary physiological pH. In addition, fungizone,

streptomycin, and penicillin were added to prevent contamination. Media requires sterile

0.2 micron filtration prior to use. Typically 3 to 5% charcoal-stripped fetal bovine serum

was used, but the best amount and the impact of serum was not fully investigated. See

Table 3-2 for individual components of the medium used for culturing largemouth bass

primary hepatocytes.




Hepatocyte Isolation

Two essential steps in a largemouth bass primary hepatocyte culture experiment

are the isolation of a homogenous population of hepatocytes and the culturing of these

isolated cell types. The latter step will be discussed in the next section. The first step is

dissociating all cell types in the liver, then selectively removing viable hepatocytes. The

other cell types include biliary epithelial cells, perisinusoidal fat-storing cells of Ito,








melano-macrophages, sinusoidal endothelial cells, and red blood cells. This first

isolation step is very technically challenging and is dependent on a healthy fish liver and

successful cannulation of the portal vein.


The Fish

Largemouth bass may be distributed nationally and live in most bodies of

freshwater, but they have yet to be successfully reared in tanks. In fact, this fish species

has to be raised on artificial feed from birth in order to control the diet. Largemouth bass

raised on artificial feed are ideal since diet may have an impact on the interpretation of

experimental results. For primary fish hepatocytes it is important to have a consistent

and reliable source of fish. For long-term maintenance, LMB should be kept in small

ponds and fed artificial feed. A small population can be kept in tanks for easy access, but

only for short periods of time, since it has not yet been demonstrated that LMB can

survive long-term in tanks. For purposes of this research, juvenile or male fish were used

for experiments in estrogen-regulated gene expression to minimize host female-specific

steroid influences. Obviously larger fish have larger livers and that usually results in a

larger yield of hepatocytes per experiment. Even though small LMB (100 g) have

smaller vessels, they are suitable to work with, and should provide successful hepatocyte

isolation.


The Perfusion

Primary hepatocyte isolation involves two-step perfusion of the liver by

cannulating the hepatic portal vein with retrograde flow out of the heart (Figure 3-1).

The objective of the particular technique is to utilize the blood vessel distribution of the

liver to administer the enzyme collagenase. This enzyme delivered in solution will








dissociate cell-to-cell interactions of the organ, resulting in a cell suspension within the

organ "pouch". Once the cells are dissociated, hepatocytes can be differentially selected

and recovered in culture.

Following proper anesthetic (150 ppm MS-222), the fish is injected intravenously

with heparin (10,000 U/Kg) to prevent blood clotting during the procedure (LMB have a

very active clotting mechanism). The fish is opened with a long ventral incision from

anus up to the mouth. Following lateral incisions on both sides, the ventral skin is

removed or pinned back to expose the viscera. Because of the importance of sex

determination, the gonads are removed for quick microscopic verification. Connective

tissue surrounding the liver is removed, exposing the gall bladder and hepatic portal vein.

First the bile is removed from the gall bladder using a lcc tuberculin syringe in order to

prevent contamination of liver cells. Then two loose square knots are tied around the

hepatic portal vein with silk sutures (3-0). Holding the hepatic portal vein with tweezers

a couple of centimeters posterior to organ entry, an IV-cannulae is inserted carefully into

the vessel. Once in, the cannulae is slowly advanced toward the organ while removing

the needle. Just prior to organ entry, the cannulae is secured in place using the two

prelooped silk sutures. Then the cannulae is attached to IV extension tubing which is

attached to a flask of solution 1 containing oxygen via a peristaltic pump (Figure 3-1).

Once all the tubing is connected, the pump is started very slowly, and solution movement

is monitored very closely. If there is no leakage by the cannulae and the organ bloats

upon solution entry, then the heart is quickly clipped to allow for release of pressure.

Solution 1 is pumped slowly (2 mL/min) through the organ in situ for 5 to 10 min in

order to remove all traces of blood from the organ. The liver should blanch in color as