Zebrafish As a Model for Studying Estrogen Signaling and Estrogenic Endocrine Disruption

Home , Bisphenol F, KLF3, OPN1LW

Zebrafish as a model for studying estrogen signaling and estrogenic endocrine disruption

______

A Dissertation

Presented to

The Faculty of the Department of Biology and Biochemistry

University of Houston

______

In Partial Fulfillment

Of the Requirements for the Degree of

Doctor of Philosophy

______

Ruixin Hao

August, 2013

Zebrafish as a model for studying estrogen signaling and estrogenic endocrine disruption

______Ruixin Hao

APPROVED:

______Dr. Jan-Åke Gustafsson, Ph.D. Committee Chair

______Dr. Maria Bondesson Bolin, Ph.D.

______Dr. Cecilia Williams, Ph.D.

______Dr. Amy Sater, Ph.D.

______Dr. Daniel Wagner, Ph.D.

______Dr. Dan Wells, Ph.D. Dean, College of Natural Sciences and Mathematics

Acknowledgements

I would like to thank Dr. Gustafsson for giving the great opportunity to work with all the

excellent scientists in our lab, giving me a very interesting project, and giving me the

guidance during my graduate studies at University of Houston.

I thank Dr. Maria Bondesson for her patient guidance and assistance all these years. She

took care of me like her own child. Without her, I wouldn’t get this far.

I thank the members of my committee members Dr. Cecilia Williams, Dr. Amy Sater, and Dr. Daniel Wagner for their time and effort during my graduate studies.

I thank my lab members Dr. Anne Riu, Dr. Nicole Durcharme, Dr. Catherine McCollum,

Caroline Pinto, Sharanya Maanasi Kalasekar, and Triet Truong for all their help and

support.

I thank our collaborators Dr. Amar Singh and Dr. Tom Knudson from EPA, Dr. Daniel

Gorelick from the Carnegie Institution for Science, Dr. Patrick Balaguer from IRCM,

France, and Dr. Anders Berkenstam and Dr. Carly Filgueira from TMHRI , Houston for

their great contribution in our collaboration.

I thank the professors and students in our center, who have given me valuable advice and

taught me experiments during my graduate studies: Dr. Margaret Warner, Dr. Chin-Yo

Lin, Dr. Daniel Frigo, Dr. Anders Ström, Dr. Christoforos Thomas, Dr. Shaun Zhang, Dr.

Weihua Zhang, Dr. Ka Liu, Dr. Yubing Dai, Dr. Laure Maneix, Dr. Rodrigo Barros, Dr.

Wanfu Wu, Dr. Lena Herbst, Dr. Bo Huang, Dr. Xingping Fu, Dr. Armando Rivera,

iii

Trang Vu, Philip Jonsson, Anne Katchy, Jun Wang, Nicholes Candelaria, Kaberi Das, Jie

Han, Lihua Tao, Jayantha Tennakoon, Effrosyni Cuko, Christopher Brooks, Prasenjit

Dey, Fotis Nikolos and Gayani Rajapaksa. I also thank all the other members in our

center for composing a happy working environment and being part of my life.

I would also like to thank my family and friends for supporting me all along all through these years. No matter how hard it was, they were always there for me. Special thanks to my parents, Liangchi Hao, Jingjng Yu, Ryan Butler, and Xin Wang for their support and help during these years.

Lastly, I would like to thank our funding EPA grant number R834289 and Robert A.

Welch Foundation.

Zebrafish as a model for studying estrogen signaling and estrogenic endocrine disruption

______

An Abstract of a Dissertation

Presented to

The Faculty of the Department of Biology and Biochemistry

University of Houston

______

In Partial Fulfillment

Of the Requirements for the Degree of

Doctor of Philosophy

______

Ruixin Hao

August, 2013

Abstract

Estrogen signaling, mediated by estrogen receptors (ERs) and G-protein-coupled estrogen

receptor (GPER), plays important roles in humans and wildlife. Perturbing estrogen

signaling may lead to deleterious health problems such as infertility, developmental

defects, metabolic disorders, and certain types of cancers. A broad range of chemicals,

which are classified as estrogenic endocrine disruptors (EEDs), can bind to ERs or

GPERs, thereby interfering with estrogen signaling and causing adverse effects.

Zebrafish embryos/larvae are an emerging model used in xenoestrogen studies. However, endogenous estrogen signaling in developing zebrafish embryos/larvae is still poorly understood. In the studies of this dissertation, first a whole-genome analysis of estrogen regulated genes in zebrafish embryos/larvae at different developmental stages was performed. The identified estrogen-responsive genes were distinct between the four time points, however GO biological process enrichment of the estrogen-responsive genes revealed similar functional groups between the four time points. Tissue-specific effects of estrogen were also analyzed using tissue enrichment of estrogen-responsive genes and

Tg(5xERE:GFP) transgenic fish. Brain, liver, heart, and pancreas were major estrogen- responsive organs in developing embryos and adults. Several candidate biomarkers were suggested from our studies, including f13a1a, cpn1, and zp3. Second, estrogenic and anti- estrogenic effects of ten bisphenols were assessed using zebrafish larvae. Selected estrogen-responsive genes identified from transcriptome analysis were tested in wild-type zebrafish larvae exposed to these bisphenols. Transgenic reporter fish Tg(5xERE:GFP)

were used to study the tissue specific effects of the selected bisphenols and GFP

vi quantification was used to compare the estrogenic effects of the bisphenols. Anti- estrogenic effects of the bisphenols were also studied in zebrafish larvae. Eight bisphenols showed tissue-specific and dose-dependent estrogenic effects in zebrafish larvae. One bisphenol showed only anti-estrogenic effects, and one bisphenol showed neither estrogenic nor anti-estrogenic effects. f13a1a was shown to be an effective new biomarker for assessing estrogenicity of chemicals. Third, high throughput screening of

ToxCastTM Phase I pesticides library and NIH Clinical Collection drug library was performed using Tg(5xERE:GFP) reporter fish. A total of 26 environmental estrogens and 62 clinical drugs with estrogenic activities were identified from these two libraries.

Then environmental estrogens identified from reporter zebrafish assays were compared to xenoestrogen datasets from in vitro assays, which are maintained in ToxCast Database

(ToxCastDB). Seven chemicals were overlapping between our fish assays and all of the in vitro assays in ToxCastDB. In conclusion, zebrafish embryos/larvae are shown to be a good model for studying estrogen signaling and estrogenic endocrine disruption. These studies are important for understanding estrogen signaling in developing zebrafish and promoting the use of zebrafish larvae as a high-throughput screening model.

vii

Table of Contents

Chapter Page

Chapter 1 Introduction ...... 1

1.1 Estrogen signaling in humans ...... 1

1.1.1 Estrogen receptors...... 1

1.1.1.1 ERα and ERβ ...... 2

1.1.1.2 GPER...... 4

1.1.2 Estrogen receptor ligands ...... 4

1.1.3 Estrogen signaling and mechanism of action ...... 6

1.1.3.1 Genomic effects ...... 6

1.1.3.2 Non-genomic effects...... 8

1.2 Endocrine disruptors ...... 10

1.3 The zebrafish embryo as a model organism ...... 14

1.3.1 Zebrafish embryo as a model to study development ...... 14

1.3.2 Zebrafish embryo as a model for studying toxicology...... 15

1.3.3 Zebrafish as a model for drug discovery...... 17

1.4 Zebrafish as a model for studying estrogen signaling and estrogenic endocrine

disruption...... 18

1.5 Project aims and objectives ...... 22

Chapter 2 Methods and Materials ...... 24

2.1 Chemicals ...... 24 viii

2.2 Zebrafish husbandry...... 25

2.3 Chemical treatment...... 26

2.3.1 Embryo treatment for transcriptome studies ...... 26

2.3.2 Larvae treatment for bishphenol studies ...... 27

2.3.3 Adult fish treatment ...... 28

2.4 RNA extraction and cDNA synthesis ...... 28

2.5 Microarray and data analysis ...... 29

2.6 Real-time PCR ...... 31

2.7 Imaging ...... 31

2.8 GFP fluorescence quantification ...... 32

2.9 Chemical library screening ...... 32

2.10 Bioinformatics ...... 33

2.10.1 Biological function inference via pathway analysis and tissue

enrichment ...... 33

2.10.2 Comparison of estrogen-regulated gene expression between embryos

and adult male fish ...... 34

2.10.3 Comparison of estrogenic pesticides among transgenic zebrafish assay

and in vitro assays ...... 34

Chapter 3 Estrogen Signaling in Zebrafish ...... 36

3.1 Introduction...... 36

3.2 Results ...... 38

3.2.1 Distinct gene expression profiles are regulated by E2 during zebrafish

development ...... 38

3.2.2 Similar biological processes are regulated by E2 at stages during

zebrafish development ...... 44

3.2.3 Tissue-specific enrichment of E2 responsive genes in embryonic

zebrafish ...... 63

3.2.4 Global biomarkers of estrogen signaling for both adult and embryonic

zebrafish ...... 65

3.3 Discussions ...... 67

3.3.1 E2 regulated genes in a stage-dependent manner ...... 67

3.3.2 E2 regulated genes in a tissue-specific manner during development

...... 74

3.3.3 E2 regulated genes with conserved biological function during zebrafish

development ...... 79

3.3.4 E2-activated blood coagulation pathways in zebrafish ...... 83

Chapter 4 Assessing Estrogenic and Anti-estrogenic Effect of 10 Bisphenols Using

Zebrafish Larvae ...... 88

4.1 Introduction ...... 88

4.2 Results ...... 92

4.2.1 Bisphenols activate GFP expression in ERE:GFP transgenic zebrafish

in a tissue-specific and dose-dependent manner ...... 92

4.2.2 Bisphenols induce expression of estrogen responsive genes in zebrafish

larvae ...... 102

4.2.3 Anti-estrogenic effects of bisphenols on GFP-expression in

Tg:(5xERE:GFP) transgenic fish and on expression of estrogenic

biomarkers ...... 103

4.3 Discussions ...... 104

4.4 Conclusions ...... 118

Chapter 5 Application of Transgenic Reporter Zebrafish in High-throughput

Screening of Environmental Estrogens and Estrogenic Drugs ...... 120

5.1 Introduction ...... 120

5.2 Results ...... 124

5.2.1 Optimization of conditions for HTS of chemical libraries using

transgenic reporter fish ...... 124

5.2.2 HTS of ToxCastTM Phase I chemical library for environmental

estrogens ...... 126

5.2.2.1 Screening for environmental estrogens using Tg: (5xERE: GFP)

transgenic reporter fish ...... 126

5.2.2.2 Comparison of the hits among different estrogen-related assays

in ToxCastDB ...... 130

5.2.3 HTS of NIHCC chemical library for estrogenic drugs using Tg:

(5xERE: GFP) fish ...... 131

5.3 Discussions ...... 139

5.3.1 Hits from ToxCastTM Phase I chemical library (Environmental

estrogens) ...... 139

5.3.2 Hits from NIHCC (Clinical estrogens and phytoestrogens) ...... 146

5.3.3 Tg(5xERE:GFP) fish as a HTS model for screening for environmental

and estrogenic drugs ...... 148

Chapter 6 Conclusions and Future Perspectives ...... 150

References ...... 155

xii

List of Tables

Table Page

Table 3.1 Numbers of differentially expressed genes at different developmental stages ...... 42

Table 3.2 Top 15 up- and down-regulated transcripts at 1 dpf upon E2 treatment

(E2 vs control) ...... 45

Table 3.3 Top 15 up- and down-regulated transcripts at 2 dpf upon E2 treatment

(E2 vs control) ...... 46

Table 3.4 Top 15 up- and down-regulated transcripts at 3 dpf upon E2 treatment

(E2 vs control) ...... 47

Table 3.5 Top 15 up- and down-regulated transcripts at 4 dpf upon E2 treatment

(E2 vs control) ...... 48

Table 3.6 Co-regulated estrogen responsive genes at 1 dpf, 2 dpf, 3 dpf and 4 dpf upon E2 treatment ...... 51

Table 3.7 Primer sequences used for the RT-qPCR validation of estrogen responsive genes from the microarray ...... 52

Table 3.8 Gene ontology biological process functional group enrichment ...... 55

Table 3.9 GO terms sub-grouped into the metabolic process category ...... 56

Table 3.10 GO terms sub-grouped into the transport category ...... 57

Table 3.11 GO terms sub-grouped into the signaling pathways category ...... 58

xiii

Table 3.12 GO terms sub-grouped into the multicellular organismal development category ...... 59

Table 3.13 GO terms sub-grouped into the response to chemical stimulus

category ...... 60

Table 3.14 Number of genes enriched by the NIH-DAVID tissue enrichment

platform ...... 61

Table 3.15 ZFIN anatomy functional chart of E2 responsive genes enriched by NIH

DAVID analysis tool ...... 62

Table 4.1 Bisphenols for the study ...... 94

Table 4.2 GFP fluorescence expression patterns in Tg(5xERE:GFP) fish after

bisphenol treatments...... 95

Table 5.1 Hits from ToxCastTM Phase I chemical library ...... 129

Table 5.2 Common hits between reporter zebrafish assays and in vitro assays in

ToxCastDB ...... 135

Table 5.3 Cherry picked compounds and GFP expression patterns in the reporter

fish ...... 136

Table 5.4 Estrogenic drugs I: endogenous estrogen, steroids and steroidgenesis

related drugs ...... 140

Table 5.5 Estrogenic drugs II: Natural products ...... 141

Table 5.6 Estrogenic drugs III: Synthetic drugs ...... 142

xiv

List of Figures

Figure Page

Figure 1.1 Gene structure of ERα and ERβ ...... 3

Figure 1.2 A representation of the genomic and non-genomic effects of estrogen receptor signaling pathways ...... 11

Figure 3.1 Dose-response curves of vtg1 and esr1 expression in zebrafish embryos.

...... 40

Figure 3.2 Principle components analysis of microarray samples ...... 41

Figure 3.3 Distinct sets of genes are regulated by E2 during different times of zebrafish development ...... 49

Figure 3.4 Clustering of geneexpression profiles of E2 and vehicle treatment groups at different time points ...... 50

Figure 3.5 Comparison of E2 regulated genes analyzed by microarray or RT- qPCR...... 53

Figure 3.6 Maternal transfer of GFP expression in Tg(5xERE:GFP) transgenic fish at 5 hpf in the absence of E2...... 69

Figure 3.7 E2-induced GFP expression in Tg(5xERE:GFP) transgenic fish during development...... 70

Figure 3.8 Venn diagrams showing overlapping estrogen responsive genes among embryos and male adult fish ...... 71

Figure 3.9 Validation of co-regulated differentially expressed genes in embryos and adult males using RT-qPCR...... 72

Figure 3.10 GFP expression in the heart, intestine, gall bladder, ovary and testis of

E2 treated adult Tg(5xERE:GFP) fish...... 73

Figure 3.11 Postulated pathway of E2 regulation of blood coagulation during zebrafish development...... 87

Figure 4.1 EE2 induced GFP fluorescence expression pattern in Tg(5xERE:GFP) fish larvae...... 96

Figure 4.2 Comparison of the GFP expression patterns in Tg(5xERE:GFP) fish after exposure to EE2, BPA, BPAF, BPE, BPF and BPB...... 97

Figure 4.3 Comparison of the GFP expression patterns between 10 µM and 1 µM

BPF treatment ...... 98

Figure 4.4 Comparison of the GFP expression patterns in Tg(5xERE:GFP) fish after exposure to BPC, BPC(Cl) and BPAP...... 100

Figure 4.5 Bisphenol-induced GFP fluorescence quantification ...... 101

Figure 4.6 EE2 induced vtg1, vtg3, cyp19a1b, f13a1a, and esr1 expression in a dose- dependent manner (A) and the effects were mediated by ERs ...... 105

Figure 4.7 Dose-response analysis of vtg1, vtg3, cyp19a1b, f13a1a, and esr1 expression in 6 dpf DZ larvae exposed to BPA, BPAF, BPE, BPF, and BPB . . . . 107

Figure 4.8 Dose-response analysis of vtg1, vtg3, cyp19a1b, f13a1a, and esr1 expression in 6 dpf DZ larvae exposed to BPC, BPC(Cl), BPAP, BADGE, and

BPS ...... 109

xvi

Figure 4.9 GFP fluorescence quantification of Tg(5xERE:GFP) fish larvae following

bisphenol treatment in the presence of EE2...... 111

Figure 4.10 Anti-estrogenic effects of the 10 bisphenols on expression of vtg1 (A) and

vtg3 (B)...... 112

Figure 5.1 E2 dose-response curves at different exposure windows ...... 128

Figure 5.2 Venn diagrams showing overlapping environmental estrogens between reporter zebrafish assays and in vitro assays in ToxCastDB ...... 133

Figure 5.3 NIH Clinical Collection chemical library components ...... 134

Figure 5.4 GFP expression patterns of 5dpf old Tg(5xERE:GFP)zebrafish larvae exposed to 1 µM SERMs...... 145

Figure 5.5 GFP fluorescence of Tg(5xERE: GFP) zebrafish larvae exposed to EE2,

isoquercitrin and RU24969...... 144

xvii

Chapter 1. Introduction

1.1 Estrogen signaling in humans

Estrogens belong to the family of steroid hormones, derivatives of cholesterol. They play

important roles in the human body, including embryonic development, sexual

differentiation, reproduction, and metabolism, as well as regulation of the cardiovascular

and the central nervous systems. Elevated estrogen levels or deficiency of estrogens may

cause many serious health problems such as breast cancer, endometrial cancer, infertility, and osteoporosis (reviewed in [1]). The endogenous estrogens include estrone (E1), 17β- estradiol (E2), and estriol (E3), among which E2 is the most predominant and biologically activate estrogen. Production of estrogens mainly occurs in the ovaries and certain extra-gonadal tissues in women before menopause, certain extra-gonadal tissues in post-menopausal women, and testes and certain extra-gonadal tissues in men. The extra-gonadal tissues include brain, adipose tissue, bone, and vasculature in both women and men, and breast in women (reviewed in [2]). Besides endogenous estrogens, some plant compounds, synthetic chemicals, and environmental chemicals can mimic endogenous estrogens and activate estrogen signaling in the body; these are called xenoestrogens.

1.1.1 Estrogen receptors

In humans, estrogen signaling is mediated by estrogen receptor α (ERα, NR3A1, encoded by ESR1 on chromosome 6) and estrogen receptor β (ERβ, NR3A2, encoded by ESR2 on

chromosome 14) (reviewed in [3]), and a membrane G-protein-coupled estrogen receptor 1

(GPER) (reviewed in [4]). Estrogens can bind to these different estrogen receptors, thereby activating distinct signaling pathways in various tissues and cell types (reviewed in [3,4]).

1.1.1.1 ERα and ERβ

Similar to other nuclear receptors (NRs), both ERα and ERβ comprise five domains

(Figure 1.1, reviewed in in [5,6]). The A/B domain, located in the N-terminus, contains the activation function 1 (AF1) domain, which can interact with co-regulators in a ligand- independent manner. The C domain is the DNA binding domain (DBD), which binds to a specific DNA response element in the promoter of target genes. The D domain is the hinge domain connecting the DBD with the ligand-binding domain (LBD), playing a role in intracellular trafficking and subcellular distribution. The E/F domain located in the C- terminus contains the LBD, which is involved in ligand binding, ER dimerization, and recruitment of co-regulators via the activation function 2 (AF2) subdomain in a ligand- dependent manner. ERα and ERβ share a high sequence identity except in the N-terminal domain (Figure 1.1). In humans, there are multiple splice variants of ERα and ERβ, each with distinct roles in different tissues and cell types (reviewed in [6,7]).

ERs are widely expressed in various tissues throughout the body, but in some tissues, one

ER subtype is dominant. ERα is mainly expressed in the liver, brain, vascular system, adipose tissue, bone, testes, epididymis, prostate (stroma), uterus, ovary (theca cells), and mammary gland. ERβ is mainly expressed in the brain, lung, colon, vascular system,

Ligand independent DBD LBD

1 184 263 302 553 595

ERα N A/B C D E F C 67kD

1 148 214 304 500 530

ERβ N A/B C D E F C 59kD

AF-1 AF-2

Homology 24% 98% 30% 59%

Figure 1.1 Gene structures of ERα and ERβ. Both ERα and ERβ are comprised of A/B domain, C domain (DNA binding domain), D domain (hinge domain), E domain (ligand binding domain), and F domain.

bladder, breast, adipose tissue, immune system, prostate (epithelium), and ovary

(granulosa cells) (reviewed in [3,8-10]). However, the expression levels of the two ER

subtypes vary in different tissues and in different situations, for example, during

development, aging, or disease. In addition, their expression levels are cyclic depending

on the presence of their ligands (reviewed in [9]). The ratio of ERα and ERβ expression

determines the estrogenic effects in different tissues and cell types (reviewed in [3]).

1.1.1.2 GPER

GPER (also known as GPR30), a member of the G-protein-coupled receptor superfamily,

is a seven trans-membrane receptor. It has been cloned from various tissues such as intestine, CNS, lung, heart, lymphoid tissue, and breast cancer cells [11-16]. GPER

mediates rapid E2 signaling by activating extracellular signal-regulated kinases (ERKs)

[17], generating cyclic adenosine monophosphate (cAMP), and inducing expression of

Bcl-2, nerve growth factor, and cyclin D2 [18-21]. GPER is involved in the regulation of

various physiological responses including functions in the reproductive, nervous, immune,

cardiovascular and renal systems, pancreas, bone, and chondrocytes (reviewed in [4]).

Abnormal expression of GPER correlates with certain malignancies such as breast,

endometrial, and ovarian cancer (reviewed in [4]).

1.1.2 Estrogen receptor ligands

Unlike the ligand-binding domain of some nuclear receptors such as thyroid hormone

receptors (TRs), whose ligand-binding cavities are well adapted to fit the cognate

hormone, ER ligand-binding domains are generous for a wide range of compounds

(reviewed in [6]). Besides endogenous estrogens, many compounds with diverse structures are able to fit in the ligand-binding cavities of ERs. These compounds include phytoestrogens (natural plant estrogens), synthetic estrogens and environmental estrogens, which are classified as endocrine disrupting chemicals (EDCs).

According to their functions, ER ligands can be categorized as agonists, antagonists, and selective estrogen receptor modulators (SERMs) (reviewed in [3,6,9,22]). ER agonists are ligands that bind to ERs and activate estrogen signaling, while estrogen antagonists are ligands that bind to ERs but block the activation of the estrogenic effects by the agonists

(reviewed in [3,22]). Agonists that can fully mediate estrogenic effects in cell systems are known as full agonists, such as E2 and ethinylestradiol (EE2). There are compounds known as partial agonists, which are able to activate ERs but not as strongly as full agonists, such as genistein (GEN) (reviewed in [22]). SERMs include two classes of chemicals, classified depending on their actions on the two ER subtypes (reviewed in [9]).

The classical SERMs (also called non-subtype-selective SERMs) can bind to both ERα and ERβ with similar potency, but differ in efficacy; examples include tamoxifen and raloxifene ([23], reviewed in [24]). Another class of SERMs encompasses those that exclusively or preferentially bind to only one ER subtype. For example, a selective ERα agonist is 4,4',4''-(4-Propyl-[1H]-pyrazole-1,3,5-triyl) trisphenol (PPT) and a selective

ERβ agonist is 2,3-bis(4-Hydroxyphenyl)-propionitrile (DPN) ([25,26], reviewed in [3]).

However, both classes of SERMs act in different ways from the endogenous estrogens and may have tissue-specific effects (reviewed in [9]).

The molecular mechanisms of ER agonists and antagonists are complex in organisms.

Upon ligand binding, the LBDs of ERs undergo conformational changes, resulting in the

recruitment of different co-regulator complexes, thereby causing different estrogenic

effects. A short helical region, Helix 12 (H12), located in the C-terminus of the LBD, is

crucial for the conformational change of the LBD (reviewed in [6]). Different ligands

cause different orientations of the H12 thereby influencing the conformation of the LBD.

When ER agonists bind to the LBD, a transcriptionally active conformation is formed.

H12 is positioned across the entrance to the ligand-binding pocket and a shallow

hydrophobic binding site on the LBD is formed as a docking surface for the leucine-rich

LxxLL motifs of NR co-activators [27-29]. The mechanisms of ER antagonists are more

complex. One of the several possible mechanisms is that some antagonists trigger a

different conformational change in the LBD of the ERs by interfering with the H12. The

antagonist-ER complex then recruits co-repressors to the target gene promoters and inhibits the gene expression ([27,30], reviewed in [6]).

1.1.3 Estrogen signaling and mechanism of action

Estrogens exert some of their effects through the ERs by directly regulating gene

expression (genomic effects); however, some effects are through rapid, protein-kinase

cascades (non-genomic effects). In addition, GPER can also mediate non-genomic

regulation of estrogens in the cellular context.

1.1.3.1 Genomic effects

According to the canonical model, ERs function in the nucleus and cause genomic effects.

In the absence of ligands, ERs are complexed together with heat shock proteins (HSP) in

the cytoplasm. Upon ligand binding, ERs dissociate from HSP and translocate into the cell nucleus. After dimerization, ERs bind to the estrogen response element (ERE) of the promoter of the target genes and recruit a co-regulator complex, thereby initiating or repressing transcription of the target genes (Figure 1.2, reviewed in [31]). ERs can form homo-dimers (ERα-ERα or ERβ-ERβ), hetero-dimers (ERα-ERβ) or a combination of the two, which is determined by the characteristics and concentration of the ligands present and the number of the receptors present. ERβ selective ligands promote the formation of

ERβ homo-dimers only. ERα selective chemicals promote the formation of both ERα-

ERα homo-dimers and ERα-ERβ hetero-dimers [32]. Ligand-dependent formation of the different dimers results in different patterns of gene regulation [33]. In engineered cell lines expressing ERα or ERβ, both ERs share similar chromatin-binding sites such as the estrogen response element (ERE). However, when both ERs are present in the cells, ERα preferentially binds to the ERE, and displaces ERβ to new binding sites enriched for other transcription factor DNA binding motifs [34].

Besides the classic genomic action model, several other models for estrogen action have been proposed. First, ERs can bind to other transcription factors, such as activator protein

1 (AP1) and specificity protein 1 (SP1), thereby initiating transcription of a different set of target genes whose regulatory sequences do not harbor an ERE (reviewed in [35]).

This is known as ERE-independent genomic action. In addition, instead of being 7

activated by ligands, ERs might be activated by phosphorylation mediated by growth

factor (GF) activated protein kinase cascades, thereby translocating into the nucleus and

initiating transcription (reviewed in [31]). In summary, ERs mediate the genomic action

of estrogen signaling by regulating gene transcription through direct or indirect binding to

the regulatory sequences of the estrogen responsive genes (ERG) in a ligand-dependent

or independent manner.

1.1.3.2 Non-genomic effects

In the past few years, research on rapid non-genomic effects (from seconds to minutes) of

estrogens has emerged, including activation of protein kinase pathways, activation of

phospholipase C, and regulation of cellular second messengers (potassium, calcium,

cAMP, and nitric oxide) (reviewed in [9,31,36,37]). E2 activates the mitogen-activated

protein kinase (MAPK) signaling pathway in different cell types, including heart [38,39], neuroblastoma [40], bone [41], endothelial [42], and breast cancer cells [43,44]. In addition, E2 can activate the phosphoinositide 3-kinase (PI3K) signaling pathway in liver

[45], endothelial [42], and breast cancer cells [46]. Several teams have reported that ERs also localize to the cell membrane in addition to the cytoplasm and nucleus [47-49].

Acconcia and colleagues reported that ERα could undergo palmitoylation and interact with caveolae rafts in the cell membrane [50]. According to Lu and colleagues, striatin acts as the molecular anchor to localize ERα to the membrane. Then an ERα-Gαi complex is formed to mediate the E2 activated MAPK, PI3K and endothelial NO synthase (eNOS) pathways [38,39]. GPER also mediates non-genomic action of the estrogens. Activation

8 of GPER by E2 stimulates production of cAMP [51], intracellular calcium mobilization

[16] and PI3K activation ([16], reviewed in [4]).

Besides E2, some SERMs and estrogen-dendrimer conjugates can also cause non- genomic effects. For example, the ERβ selective agonist DPN, but not the ERα selective agonist PPT, increases intracellular calcium within a few minutes of treatment, and activates the MAPK and PI3K/AKT pathways in neuronal cells [52]. Estrogen-dendrimer conjugates can exclusively activate cytoplasmic ER signaling but not ER-mediated nuclear events [44,53]. In MCF-7 breast cancer cells, they stimulate the phosphorylation of extracellular signal-related kinase (ERK), Rous sarcoma oncogene (Src) and Src homology 2 domain containing transforming protein (Shc), but are less effective in stimulating estrogen target gene transcription or breast cancer cell proliferation [53].

However, genomic regulation and non-genomic regulation of E2 might come to a convergence, due to the complexity of gene regulation. The promoters of the target genes might be regulated by more than one transcription factor that might be activated by different signaling pathways through both genomic and non-genomic actions. For example, the c-fos gene is regulated by ERs through both genomic and non-genomic actions. Duan and colleagues reported that c-fos can be activated by the interaction of

ERs with the Sp1 transcription factor at its GC-rich promoter sequences in the presence of E2 [54]. Later the same group found that c-fos can also be activated by Elk-1 and SRE transcription factors, through activation of MAPK and PI3K signaling pathways upon E2 treatment in MCF-7 breast cancer cells [55,56]. Moreover, Cyclin D1 and LDL-R genes are also activated by E2 through both genomic and non-genomic actions [45,46,57-59]. 9

In summary, estrogen regulation is a complex network, including both genomic and non-

genomic actions, which eventually may converge in the regulation of specific target

genes (Figure 1.2).

1.2 Endocrine disruptors

The term endocrine disruptors (EDs, also called endocrine disrupting chemicals, EDCs)

was first brought up by Dr. Ana Soto and her collaborators in 1993 [60]. It refers to

compounds that have the potential to disturb the normal endocrine system by binding to hormone receptors, thereby causing adverse health problems in humans (reviewed in

[61,62]). A wide range of chemicals have been reported to be endocrine disruptors, including phytoestrogens, dioxins, pesticides, organotins, and components of plastics

(such as polyfluoroalkyl compounds, brominated flame retardants, alkylphenols, bisphenols, and phthalates) (reviewed in [62]). As the global production of these chemicals in industry increases, pollution of numerous media like air, water, and food is increasing. More and more exposure occurs and the mixed exposure of these EDCs may cause severe health problems.

There is emerging evidence for deleterious effects caused by exposure to EDCs. These effects include developmental defects, infertility, metabolic disorders, obesity, and certain types of cancer (reviewed in [60-62]). EDCs can act by interfering with the

production, release, metabolism and elimination of the natural hormones and mimicking

their effects in the body (reviewed in [63]). Various hormone systems can be targeted by

Membrane receptors Ligands

Protein kinase

Non-genomic effects ER

Cytosol Nucleus

Co-regulator RNA pol Or 2 complex ×

ER ER mRNA

ERE Target gene

Gene transcription

Figure 1.2. A representation of the genomic and non-genomic effects of estrogen receptor signaling pathways.

EDCs, such as estrogen, androgen, thyroid, and glucocorticoid systems (reviewed in [62]).

Exposure dosage, exposure window, and the complexity of the EDCs present in the

environment all determine the different endpoints of the adverse effects. Differences

between low-dose and high-dose effects of EDCs have been controversial in this field

(reviewed in [64-66]). Exposure window is also crucial for the toxic effects. For example,

if a pregnant female is exposed to the EDCs, the germ cells might be directly affected and

the adverse effects might be transgenerational [67]. The pharmaceutical estrogen

diethylstilbesterol (DES) causes a multigenerational phenotype. Exposure of a pregnant

female mouse to DES not only causes reproductive tract abnormalities and gonadal

dysfunction in the F1 generation males and females, but also causes abnormal female

reproductive tract function in the F2 generation in mice [68].

Many mechanisms have been proposed for endocrine disruption (reviewed in [62]). Some

EDCs act as direct agonists or antagonists to nuclear receptors (NRs) by competing with

the endogenous hormones and recruiting related co-activator or co-repressor complexes.

Some EDCs bind to NRs and interfere with other hormone receptor signaling by

competing for the common co-regulators or common DNA-binding sites. In some other

cases, proteasome-mediated degradation might occur upon the EDC interactions with the

NRs (reviewed in [62]). EDCs can also act through altering the epigenome, thereby

causing long-term, even transgenerational effects [69].

Estrogenic endocrine disruptors (EEDs, also called xenoestrogens) are compounds that

bind to the ERs and mimic the natural estrogen effects in the body (reviewed in [61,70]).

The sources of EEDs include plants (e.g. phytoestrogens), synthetic pharmaceuticals (e.g. 12 ethinyl-estradiol), pesticides (e.g. methoxychlor), and industry materials (e.g. bisphenol

A) (reviewed in [70]). The synthetic estrogen EE2 used in contraceptive pills has been detected in surface water and sewage effluent outflow. The detected concentration is within the range that may cause long-term exposure effects in humans and wildlife

[71,72]. Methoxychlor (MTX) was used as a substitute for dichlorodiphenyltrichloroethane (DDT) in pesticides since DDT was banned because of its endocrine disruption activities. Later MTX was also reported as an EED ([73], reviewed in [70]). Moreover, its major metabolite 2, 2-bis(p-hydroxyphenyl)-1,1,1- trichloroethane (HPTE) has been reported to be an agonist for ERα and an antagonist for

ERβ [74]. Therefore, the United States Environmental Protection Agency (EPA) banned the use of MTX as a pesticide. Bisphenol A (BPA), a monomer of polycarbonate plastics, has also been extensively reported as an EDC, especially as an EED (reviewed in [66]).

In adult urine, BPA has been detected in the nanomolar range (reviewed in [75]) and in babies its levels may be higher since babies do not have a fully developed metabolic system. Therefore, the usage of BPA in baby bottles has been banned in many countries such as the US, Canada, Denmark, and Germany, and more countries have started to become aware of the deleterious effects of BPA.

Interference of estrogen signaling by EEDs is mediated by both ERα and ERβ through genomic and non-genomic actions. Some EEDs may act as SERMs by selectively modulating different ER subtypes (reviewed in [61,70]). Furthermore, some EEDs can bind to aryl hydrocarbon receptor (AhR) and indirectly disturb estrogen signaling through the crosstalk of AhR and ER signaling pathways (reviewed in [70]). However, since

EEDs can target other hormone systems, the effects caused by EED exposure are broader

than just estrogenic effects.

1.3 The zebrafish embryo as a model organism

The zebrafish, Danio rerio, has become an emerging model for studying vertebrate

development, disease, organ function, behavior, genomics, toxicology, and for drug discoveries. Features like high fecundity and ex-utero, transparent, and rapid embryonic

development make them a low-cost but dynamic model. Furthermore, the fully sequenced

genome and the amenability for forward genetic screens and reverse genetics promote the

usage of this model for studying gene functions in developmental and disease processes

(reviewed in [76]). The small size and high number of offspring facilitate the use of

zebrafish embryos as a high-throughput model for drug screening and toxicity assessment

of environmental chemicals (reviewed in [77-80]).

1.3.1 Zebrafish embryo as a model to study development

Zebrafish have been used for studying the development of various organs. Their

developmental stages are generally categorized by hours post fertilization (hpf) as zygote

(0~0.75 hpf), cleavage (0.7-2.2 hpf), blastula (2.25-5.25 hpf), gastrula (5.25-10 hpf),

segmentation (10-24 hpf), pharyngula (24-48 hpf), and hatching (48-72 hpf). After

hatching, zebrafish are considered early larvae. The major organs including brain, muscle,

heart, gut, eyes, and ears are fully developed by 5 days post fertilization (dpf) [81].

During early embryogenesis, zebrafish organ development has a high similarity to that of

mammals in various systems, including neuronal, cardiovascular, somite, skeletal, and

muscular systems. For example, the zebrafish embryo has been used to study the

development of the liver, a major functional organ for metabolism, detoxification, and homeostasis. Liver morphogenesis of the zebrafish includes two phases: budding and growth [82]. The budding phase occurs from 24 hpf to 50 hpf, based on the mRNA expression of a liver marker using in situ hybridization of the developing liver [83]. At the growth stage, the liver changes size and shape and begins to be functional. At 96 hpf

(4 dpf), the second lobe (right lobe) starts to form and continues growing in size until 120 hpf (5 dpf). Moreover, zebrafish embryos have also been used to study pancreatic development. Similar to the mammalian pancreas, the zebrafish pancreas includes both an exocrine/ductal compartment and endocrine part comprising alpha, beta, delta, epsilon, and pancreatic polypeptide (PP) cells [84]. Zebrafish pancreatic morphogenesis initiates from the 16 somite stage, thereafter the endocrine and exocrine parts of the pancreas bud in different locations during development. The endocrine pancreas secretes insulin, glucagon, pancreatic polypeptide, ghrelin, and somatostatin, playing an important role in regulation of proliferation and differentiation ([84], reviewed in [85]).

1.3.2 Zebrafish embryo as a model for studying toxicology

In the past decade, the use of zebrafish embryos as a model for toxicity studies has steadily increased (reviewed in [80]). Embryo lethality and teratogenic effects are two important parameters for the risk assessment of chemicals (reviewed in [79,80]). Three general assessment criteria for toxicity are the lethal concentration for 50% of the population (LC50), the lowest observable adverse effect concentration (LOAEC), and the no observed adverse effect concentration (NOAEC) (reviewed in [79]). These 15

concentrations vary depending on the exposure window during zebrafish embryonic

development. To cover the entire window of organogenesis in a teratogenicity assay, the

treatment of zebrafish embryos is normally started from gastrulation stage (approximately

4-10 hpf) and morphology is assessed at or beyond 72 hpf when the zebrafish are at the

larval stage (reviewed in [79]). Other than lethality, endpoints for the early embryonic-

stage toxicity also include abnormal hatching rate, morphology, and behavior. Abnormal

morphology consists of altered body size, spinal curvature, altered head size and

formation, malformation of eye, jaw and other craniofacial structures, edema, lack of

swim bladder inflation, altered tail formation, altered pigmentation, and malformation in

the pericardial sac and yolk sac ([86], [87], and reviewed in [80]).

Besides the acute toxicity testing in environmental risk assessment, the zebrafish embryo

model is also used for studying the mechanisms of the toxic compounds and the

indication of adverse or long-term effects (reviewed in [78]). One of the applications is

toxicogenomics, which is a combination of toxicology and genomics to elucidate the

molecular mechanisms of the toxic compounds using high throughput technologies.

Differing from traditional toxicology studies, which focus on certain biomarkers or

pathways, toxicogenomics generate a high throughput molecular profiling and systematic pathway regulation network from the whole-genome upon treatment. Toxicogenomics includes transcriptomics, proteomics, and metabolomics, among which transcriptomics is used for studying the toxic effects of chemicals on gene expression patterns and for functional gene studies. Chen and colleagues exposed zebrafish embryos to retinoid acid

(RA) and 2, 3, 7, 8-tetrachlorodibenzo-p-dioxin (TCDD) and analyzed the transcriptome

16 profiles. They found that both compounds triggered a common transcriptional response associated with heart failure, but the set of altered genes were different. Through functional analysis, they found that the induction of gene Nr2F5, a member of COUP-TF family of nuclear receptors, was necessary and sufficient for the cardiotoxic response to

RA [88].

Furthermore, as an in vivo system, the zebrafish embryo can be used to conduct toxicokinetic studies, to take into account the rate of uptake, metabolism and excretion of the compounds to determine the bioavailablity and effective concentrations of the compounds (reviewed in [78]). In summary, zebrafish embryo as a model system has various applications in toxicology and pharmacy, and shows great potential in these fields.

1.3.3 Zebrafish as a model for drug discovery

The use of zebrafish in pharmaceutical research and drug discovery includes screening compounds, drug toxicity, target identification and validation, physiology-based drug discovery, quantitative structure-activity relationship (QSAR), and structure-activity relationship (SAR) studies (reviewed in [89]). Using zebrafish for drug screening takes into consideration the ADMET (absorption, distribution, metabolism, excretion, and toxicity) in the early drug development process, which might lead to avoidance of extensive late-stage animal testing [90]. With improving high-throughput screening (HTS) techniques, zebrafish embryos have become more commonly used in HTS drug screening

(reviewed in [91,92]). A system for loading zebrafish from reservoirs or multiwall plates and positioning and rotating them for high-speed confocal imaging has been developed

which greatly facilitates the use of zebrafish for HTS [93]. Automated detection and high-resolution imaging platforms were also developed for HTS of zebrafish embryos

[94-96]. The generation of transgenic reporter fish lines has greatly facilitated HTS.

Using transgenic zebrafish with fluorescent blood vessels, Tran and colleagues identified

several antiangiogenic compounds by HTS of a LOPAC1280 compound library [97].

Molina and colleagues performed a screen using Tg(dusp6:EGFP) transgenic reporter

fish and discovered a small molecule, Dusp6-(E)-2-benzylidene-3-(cyclohexylamino)-

2,3-dihydro-1H-inden-1-one (BCI), as a Dusp6 inhibitor [98]. These studies show that

the zebrafish model is promising for HTS in drug development.

1.4 Zebrafish as a model for studying estrogen signaling and estrogenic endocrine

disruption

Since estrogen signaling is highly conserved in vertebrates, zebrafish can be used as a

model for studying estrogen signaling. There are three ERs (Esr1a, Esr2a and Esr2b) in

zebrafish [99]. Esr1a is the human ERα homolog, while Esr2a and Esr2b are human ERβ homologs. The overall amino acid sequence identity between zfERs and hERs is approximately 50%; however, the DBD (~90%) and LBD (~60%) are more conserved

[100]. In addition, another report has shown that the LBD of Esr1 in zebrafish has a comparable identity with the LBD of both hERα and hERβ, but Esr2a and Esr2b of zebrafish are more identical with hERβ [101]. Zebrafish also have a fourth estrogen- targeted receptor, the membrane localized G protein-coupled estrogen receptor 1 (Gper).

Gper induces activation of a non-genomic response, including phosphorylation of the

18 mitogen-activated protein kinases MAPK3/MAPK1 [102], which eventually results in downstream transcriptional changes.

In zebrafish, expression of ER subtypes is tissue specific and time dependent. In zebrafish embryos, ER expression has been detected from 24 hpf [99,103-105]. Esr2a is more predominant in early development than Esr1 and Esr2b because of its maternal load in the oocyte. However, Esr2a disappears between 6 and 12 hpf and gradually returns with the zygotic expression [99,106]. At 24 to 48 hpf, esr2a mRNA is mainly expressed in the head/brain region and in proximity to the yolk as determined by in situ hybridization. At later stages in the early development, it is expressed mainly in the epidermis, pectoral ﬁn buds, hatching gland and neuromast cells [107-109]. mRNA expression of esr1 and esr2b has been shown to steadily increase from 6 hpf during embryonic development using real-time qPCR (RT-qPCR) [103,105]. High levels of esr2b and low levels of esr1 were also detected in the brain, epidermis, pectoral fin buds, and hatching gland of the embryos at 24 hpf as shown by in situ hybridization [104].

Morpholino knock-down of esr2a transcript increased apoptosis in embryos and caused severe malformations of 1-5 dpf larvae. Teratologic phenotypes include deformed body curvature as well as growth delay, abnormal brain, smaller eyes and otic vesicles, heart edema, delay of swim bladder development, yolk utilization with swollen yolk extension, and caudal fin defect with aberrant circular swimming[107]. This indicates a significant role of esr2a in early zebrafish development. In adult fish, the expression of the three different ERs overlaps to a large extent. Major organs of ER expression are the liver, heart, brain, testis, and intestines [103]. In adult zebrafish liver, Northern blot analysis of

ER mRNA showed that esr1 expression is up-regulated by E2, however, the expression

level of esr2b was strongly reduced while the expression level of esr2a was barely

affected. This might indicate that esr2a and esr2b have different expression patterns when physiological E2 levels fluctuate, and they may play distinct roles in regulating liver functions [110] Although ER mRNA expression patterns have been well studied,

ER signaling in zebrafish, especially during embryonic development, remains poorly understood.

Zebrafish have been used to study EEDs, with the exposure windows spanning the embryo, larva, juvenile, and adult stages (reviewed in [111]). One of the most common endpoints for the EED studies using zebrafish is to measure the expression of estrogenic

biomarkers such as Vitellogenins (Vtgs), Cytochrome P450 aromatase B (Cyp19a1b) and

Esr1. Vtgs, including Vtg1, Vtg2, Vtg3, Vtg4, Vtg5, Vtg6, and Vtg7, are a family of egg-

yolk precursor proteins that are synthesized in the zebrafish embryo and the liver of

female adult fish under the control of E2. They can, however, be induced in the liver of

embryos and male fish upon treatment with estrogen and xenoestrogens. Therefore the induction of Vtgs in males has been used as a sensitive and appropriate endpoint to

screen for EEDs [112]. It has been reported that diethylstilbestrol (DES), EE2, BPA and

genistein can all induce Vtg expression in adult zebrafish or embryos [112-116].

Cyp19a1b is involved in the process of estrogen biosynthesis by converting androgens

into estrogens via aromatization. It is expressed mainly in the brain from embryonic

stages and weakly in the gonadal tissues [117,118], and it has also been used as an

estrogenic biomarker. Its expression was reported to be induced by EEDs such as EE2,

DES, and BPA ([119], reviewed in [120]). These biomarkers are appropriate for some

EEDs but not all of them due to the complexity of estrogen signaling. Therefore, more biomarkers are needed to better assess the estrogenic effects of the EEDs.

Recently, several transgenic reporter fish lines have been generated, which greatly facilitate the screening for EEDs. Legler and colleagues generated a transgenic fish containing a luciferase reporter gene downstream of ERE (reviewed in [121]). Exposure to EEDs activates estrogenic response in the transgenic fish by stimulating the luciferase gene transcription, which is measured by luciferase activity in tissue lysates [122].

Moreover, Tg(ERE-ZVTG1-EGFP) transgenic fish [123] and Tg(cyp19a1b-GFP)

transgenic fish [124] have also been generated for EED studies. Unfortunately, the use of

these two reporter fish lines may be limited for the study of EEDs, because EEDs might

have no effect on the expression of Vtg1 and Cyp19a1b. Recently two ERE:GFP reporter

fish lines were established. Gorelick and Halpern made a transgenic reporter zebrafish

Tg(5xERE:GFP), which has 5xERE upstream of a green fluorescent protein (GFP)

reporter to study estrogen signaling in zebrafish [125]. GFP expression is activated in

estrogenic tissues such as liver, brain, heart, and ovary upon E2 treatment in these fish.

Lee and colleagues generated Tg(3×ERE:Gal4ff, UAS:GFP), which contains 3xERE and a Gal4ff-UAS system as well as GFP reporter. This transgenic fish shares some common estrogen responsive tissues with the Tg(5xERE:GFP) fish, visualized by GFP expression in organs such as the liver, heart, and brain, but exhibits additional GFP expression in skeletal muscle and lateral line upon E2 treatment [126]. These transgenic fish provide

insights into the tissue-specific effects of estrogen and show great potential for HTS of

chemicals; nevertheless, these studies are still open for improvement.

Besides being used for chemical screening and toxicity assessment, zebrafish can also be

a valuable model for studying the mechanisms of EEDs and targets of EED action.

Kallivretaki and colleagues knocked down cyp19a1b using the mopholino technique in zebrafish embryos, and noticed that the number of neuromast hair cells was decreased

during zebrafish development [127]. This study provided evidence for the role of

estrogens in the development of the lateral line mechanosensory system of zebrafish.

In summary, the zebrafish model is a valuable tool for studying estrogen signaling and

estrogen disruption. However, the mechanism of the estrogen signaling in zebrafish

remains to be further studied and the potential for zebrafish as a model for HTS needs to

be further explored.

1.5 Project aims and objectives

Our goal was to further understand estrogen signaling in zebrafish embryos and early

larvae, and use zebrafish embryos and early larvae to investigate estrogen disruption.

Specifically, our first aim was to study estrogen signaling pathways in zebrafish embryos

and early larvae. To accomplish this, a whole-genome analysis of estrogen-regulated

genes in zebrafish embryos at different developmental stages was performed.

Differentially expressed genes were analyzed using bioinformatics methods to identify

their biological functional classes. Comparison of the differentially expressed genes in

embryos was performed to identify potential estrogenic biomarkers for both embryos and 22

adult zebrafish. Tg(5xERE: GFP) reporter fish were used to study the tissue-specific

effects of E2 in zebrafish embryos, larvae and adult fish.

Our second aim was to assess the estrogenic and anti-estrogenic effects of ten bisphenols

using zebrafish larvae. Selected estrogen-responsive genes identified in our first aim were

tested in wild-type zebrafish larvae exposed to these bisphenols. Transgenic reporter fish

Tg(5xERE: GFP) were used to study the tissue-specific effects of the selected bisphenols

and GFP quantification was used to compare the estrogenic effects of the bisphenols.

Anti-estrogenic effects of the bisphenols were also studied in zebrafish larvae.

Furthermore, we aimed to develop a high through-put method to screen for

environmental estrogens from ToxCastTM Phase I pesticide library, and clinical estrogens from NIH Clinical Collection drug library using Tg(5xERE:GFP) reporter fish. Obtained

hits from ToxCastTM Phase I were compared with hits from estrogen-related in vitro

assays from ToxCast database (ToxCastDB) maintained by EPA.

Chapter 2. Methods and Materials

2.1 Chemicals

17β-estradiol (E2), 2,2-Bis (4-hydroxyphenyl) propane (BPA), 2,2-Bis (4- hydroxyphenyl) hexafluoropropane (BPAF), 2,2-Bis (3-methyl-4-hydroxyphenyl) propane (BPC), 4,4′-Isopropylidenediphenol diglycidyl ether (BADGE), 17α-ethinyl- estra-1,3,5(10)-triene-3,17β-diol (EE2), 4OH-Tamoxifen (4OH-TAM), genistein (GEN), tributylstannane (TBT), diarylpropionitrile (DPN), ICI 182,780 (fulvestrant), and dimethylsulfoxide (DMSO) were obtained from Sigma Aldrich (St. Louis, MO, USA).

1,1-Bis (4-hydroxyphenyl) ethane (BPE) was obtained from 3B Scientific Corporation

(Wuhan, China). Bis (4-hydroxyphenyl)-2,2-dichlorethylene (BPC-Cl), bis (4- hydroxydiphenyl) methane (BPF), 2,2-bis (4-hydroxyphenyl) butane (BPB), bis (4- hydroxyphenyl) sulfone (BPS), estrone (E1), estriol (E3), ERB041, WAY200070,

FERB033, propylpyrazole triol (PPT), 16alpha-lactone-estradiol (16αLE2), and 8-beta- vinyl-estradiol (8bVE2) were obtained from Patrick Balaguer at Institut de Recherche en

Cancérologie de Montpellier (IRCM, Montpellier, France), purchased from Sigma

Aldrich (Sigma Aldrich, Saint Quentin Fallavier, France). Medicarpin was obtained from

Anders Berkenstam at the Methodist Hospital Research Institute (TMHRI, Houston, TX,

USA). The chemicals were prepared in 100% DMSO at a stock concentration of 10 mM.

Two chemical libraries, ToxCastTM Phase I (ToxCast) and NIH Clinical Collection I

(NIHCC) were obtained from the EPA and TMHRI, respectively. The EPA’s Phase I

ToxCastTM library contains 320 substances and 309 unique chemicals, most of which are

food-use pesticides [128,129]. Structure Data Format (SDF) files can be downloaded at 24

http://www.epa.gov/NCCT/toxcast/chemicals.html. The environmental chemicals were

prepared in 100% DMSO at a stock concentration of 20 mM. Quality control (QC)

information for all the chemicals is available on the EPA’s ToxCast

website: http://www.epa.gov/ncct/toxcast/chemicals.html. NIHCC contains 446 small

molecules that have a history of use in human clinical trials. This library is assembled by

the National Institutes of Health (NIH), with the aim of discovering new bioactivity and

therapeutic potential for these chemicals. The clinical chemicals were prepared in 100%

DMSO at a stock concentration of 10 mM. The detailed information for all the chemicals

is available at http://www.nihclinicalcollection.com.

2.2 Zebrafish husbandry

Wild-type DZ strain and transgenic fish Tg(5×ERE:GFP) were used according to the

maintenance and experimental protocols approved by the Institutional Animal Care and

Use Committee at University of Houston (protocol numbers 12-042 and 10-040 approved

at Sept. 17, 2012 and Nov. 19, 2012, respectively). Adult zebrafish were maintained in

2.5 Liter polyethylene tanks in a Z-MODE holding system from Aquatic Habitat,

(Aquatic Habitats Inc., Apopka, FL) or 3.5 liter tanks in a Tecniplast system (Tecniplast

USA Inc., West Chester, PA) supplied continuously with circulating filtered water at

28˚C under 14 h of light and 10 h of dark cycle (14:10 LD; lights on 8 AM; lights off 10

PM). The fish were fed commercial flake food (Aquatic Habitat) in the morning, baby brine shrimp (Brine Shrimp Direct, Ogden, UT) at noon and Cyclop-eeze (Argent

Chemical Laboratory, Redmond, WA) in the evening during the week. On weekends they

were fed baby brine shrimp and Cyclop-eeze.

To collect embryos, adult zebrafish were mated in several breeding tanks on the same day,

with 4 adult fish in each breeding tank. On the second day, embryos were gathered from

each breeding tank, pooled and cultured in embryo medium (E3, 5 mM NaCl, 0.17 mM

KCl, 0.33 mM CaCl2, 0.33 mM MgSO4) in 100x15mm Petri dishes (VWR, Houston, TX)

at 28.5˚C with 14:10 LD until treatment.

2.3 Chemical treatment

2.3.1 Embryo treatment for transcriptome studies

A clutch of pooled DZ zebrafish embryos were divided and transferred into 6 well plates

(VWR, Houston, TX) with 30 embryos per well in 3 mL E3 embryo media. E2 was

prepared to a 1 mM working stock solution in 100% DMSO. 3 µL 1mM E2 was added

into 3 mL E3 to obtain a 1 μM final concentration. Treatment started at approximately 3

hours post fertilization (hpf). 0.1% DMSO (vehicle alone) was used as negative control.

Embryo media containing E2 were changed every day. At different time points, 1 day

post fertilization (dpf, 24 hpf), 2 dpf (48 hpf), 3 dpf (72 hpf) and approximately 4 dpf (4.3

dpf, 104 hpf), embryos were collected for RNA extractions.

A clutch of pooled Tg(5xERE:GFP) zebrafish embryos were divided and transferred into

6 well plates with 25 embryos per well in 2.5 mL E3 embryo media. Embryos were

exposed to 1 μM E2 following the same procedure as discussed above. To inhibit pigmentation, 1-phenyl-2-thiourea was added into the media to obtain a 0.003% final 26

concentration from 24 hpf onward. At different time points, 1 dpf, 2 dpf, 3 dpf, 4 dpf, 5 dpf and 6 dpf, fish were anesthetized with 0.04% tricaine (Sigma-Aldrich, St. Louis, MO) and imaged.

2.3.2 Larvae treatment for bishphenol studies

After spawning, DZ embryos were cultured in a 100x15mm Petri dishes in E3 embryo medium at 28.5˚C with 14:10 LD till 4 dpf. Media were renewed every day. At 4 dpf, DZ fish larvae were transferred into 6 well plates with 25 larvae per well in 2.5 mL E3 media.

Bisphenol and EE2 stocks ranging from 10-5-10-2 M were diluted 1000-fold into each

well to reach final concentrations. 0.1% DMSO was used as a negative control. Larvae

were incubated in the chemical or 0.1% DMSO until 6 dpf and media were renewed at 5

dpf. Twenty-five of the 6 dpf DZ larvae were pooled as one sample following RNA extraction.

After spawning, Tg(5xERE:GFP) embryos were cultured in a 100x15mm Petri dishes in

E3 embryo medium at 28.5˚C with 14:10 LD till 4 dpf. PTU was added in E3 media at a

0.003% final concentration to inhibit pigmentation from 24 hpf onward. At 4 dpf, larvae

were transferred to 6 well plates with 30 larvae per well in 3 mL E3 media containing

0.003% PTU. Bisphenol and EE2 stocks ranging from 10-5-10-2 M were diluted 1000-fold

into each well to reach final concentrations. 0.1% DMSO was the negative control.

Larvae were incubated in chemical or 0.1% DMSO until 6 dpf and media was renewed at

5 dpf. At 6 dpf, Tg(5xERE:GFP) larvae were transferred into 96-well U-bottom black plates for GFP quantification or to glass slides for imaging.

2.3.3 Adult fish treatment

Four six-months-old male adult DZ fish were transferred into a 1 liter beaker containing

500 µL egg water. 500 µL 1 mM E2 were added into the egg water yielding a 1 µM final

concentration. Another four six-months-old male adult DZ fish from the same tank were

exposed in 500 µL 0.1% DMSO only as a control group. Treated adults were maintained at room temperature with 14:10 LD for 48 h, then anesthetized with 0.2% tricaine and

snap frozen in liquid nitrogen following RNA extraction.

Four male and four female six-month-old adult Tg(5xERE:GFP) fish were exposed to 1

µM E2 or vehicle DMSO (0.1%) following the same procedures as discussed above.

After the treatment, fish were anesthetized with 0.2% tricaine following dissection

according to standard procedures under a dissection microscope (Leica, Buffalo Grove,

IL). Organs were rinsed with 1xPBS (VWR, Suwanee, GA) following imaging.

2.4 RNA extraction and cDNA synthesis

Total RNA from pooled embryos was extracted using Trizol (Invitrogen Corporation,

Carlsbad, CA) and RNeasy spin columns (Qiagen, Chatsworth, CA) according to the

manufacturer’s protocols. On-column DNase I (Qiagen, Chatsworth, CA) digestion was

performed to remove remaining DNA. RNA concentrations were measured with

NanoDrop 1000 spectrophotometer (Agilent Technologies, Palo Alto, CA) and RNA

integrity was analyzed with the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo

Alto, CA). cDNA synthesis was carried out using Superscript II reverse transcriptase

(Invitrogen Corporation, Carlsbad, CA) and random-hexamer primers (Promega

Corporation, Dane, WI).

Frozen adult fish from E2 treated and DMSO control groups were ground to a crude powder using pre-cooled mortars and pestles. Liquid nitrogen was added into the mortars to keep the samples frozen. Crude tissue powder was then transferred to pre-cooled 5 ml sterile centrifuge tubes (VWR, Houston, TX) containing Trizol (Invitrogen Corporation,

Carlsbad, CA). A motorized homogenizer (Kinematica Polytron PT 1200 E, Lucerne,

Switzerland) was then used to completely homogenize the fish samples. Total RNA was extracted as described above.

2.5 Microarray and data analysis

Agilent zebrafish gene expression microarray v2 (part number G2519F and AMADID

(design) 019161) was used for the microarray analysis. Experiments were performed by the Genomic and RNA Profiling Core (Baylor College of Medicine, Houston, TX). The

Genomic and RNA Profiling Core first conducted Sample Quality checks using the

Nanodrop ND-1000 and Agilent Bioanalyzer Nano chips. For labeling, the Agilent Quick

Amp Labeling Kit (for one-color) Protocol Version 6.5 was used. 50 ng of total RNA that had passed the quality check was used for the protocol as recommended by Agilent. The

Labeling Kit (Agilent p/n 5190-0442) was used along with Agilent’s RNA Spike-In Kit,

Agilent’s Hybridization Kit, and Agilent’s Wash Buffers 1 and 2. The RNA Spike-Ins was added to the sample. The sample was simultaneously amplified and Cy3 dye-labeled cRNA was generated using T7 RNA Polymerase. The cRNA was purified using Qiagen

RNeasy mini spin columns. Samples were then measured again on the Nanodrop for yield

and dye incorporation. The samples were then fragmented and 1.65 µg of sample and

hybridization mix was loaded onto each of the 4x44K Expression arrays. The slide was hybridized in Agilent Hybridization Chamber at 65°C at a 10 rpm rotation for 17 hours.

The slide was washed using the Agilent Expression Wash Buffer Set 1 and 2 as per the

Agilent protocol. Once dry, the slides were scanned with the Agilent Scanner (G2565BA) using Scanner Version C and Agilent Feature Extraction Software Version 11.0.1.1. Time points 1 and 2 dpf were performed in biological triplicates of independent pools of RNA while time points 3 and 4 dpf were performed in quadruplicates. The biological replicates

were prepared at different time points from embryo batches of different breeding pairs,

but from the same fish strain. The microarray results were submitted to Gene Expression

Omnibus database (GSE42766).

Raw data from the microarray analysis was mean-centered and quantile-normalized to

normalize gene expression distributions across the different samples. The data was then

Log2-transformed. Batch effects from the different biological replicates were removed

using Partek Genomics Suite v 6.3 (http://www.partek.com/) and residual variance was

analyzed by Principal Components Analysis (PCA). Then the data was subjected to two-

way ANOVA to study the effect of the developmental stages, treatment and their

interactions. The development stages had the maximum effect on the gene expression,

hence one-way ANOVA (p≤0.01) was used to identify genes altered by treatments at

individual developmental stages. Venn diagrams were generated to illustrate the

overlapping genes among the four different time points (p≤0.01, absolute fold change

≥|±1.4|). For the Hierarchical clustering of figure 3, unsupervised hierarchical clustering

was performed using Pearson correlation algorithm for the gene tree and Spearman for

the developmental stages (p≤0.005).

2.6 Real-time qPCR

RT-qPCR was performed using a 7500 Fast Real-Time PCR (Applied Biosystems, Foster

City, CA) with Fast SYBRGreen Master mix (Applied Biosystems, Foster City, CA).

Primer BLAST (http://www.ncbi.nlm.nih.gov/tools/primer-blast/) was used to design the primers, which were synthesized by Integrated DNA Technologies, Inc (San Diego, CA).

Primers of 18-22 base pairs (bp) were designed to amplify sequences of 100-300 bp.

Relative gene expression data was normalized against 18S ribosomal RNA (18S rRNA) expression and analyzed with unpaired two-tailed t-test. Each experiment was carried out with at least two independent batches of embryos and three technical replicates for each

PCR reaction. Significance is presented at p≤0.05 (*) or p≤0.01 (**).

2.7 Imaging

For live imaging, embryos and larvae were incubated in 0.003% 1-phenyl-2-thiourea

from 24 hpf till the end of chemical treatment to inhibit pigment formation (Sigma-

Aldrich, St. Louis, MO). Before imaging, embryos and larvae were anesthetized with

0.04% tricaine (Sigma-Aldrich, St. Louis, MO) and mounted in 3% methylcellulose on a

glass slide. E2-treated developing embryos and larvae were visualized using a Nikon

AZ100M microscope equipped with Nikon DS digital camera head and the NIS Elements

imaging software (Nikon Instruments Inc, Melville, NY). All the other chemical treated

larvae and adult organs were imaged using an Olympus fluorescence microscope

(OLYMPUS IX51, Center Valley, PA) equipped with an OLYMPUS XM10 camera

(Center Valley, PA) and the MicrosuiteTM 5 software (Center Valley, PA). Fluorescent

images of live embryos were pseudo-colored and superimposed using Adobe Photoshop

CS5 (Adobe Systems Inc. Sam Jose, CA). Adjustments and cropping were performed

using Photoshop CS5 (Adobe Systems Inc. San Jose, CA).

2.8 GFP fluorescence quantification

For assessing estrogenic and anti-estrogenic effects of bisphenols, 6 dpf Tg(5×ERE:GFP)

larvae were anesthetized using 0.002% eugenol and transferred into 96-well U-shape

bottom black microplates (Greiner Bio-One, Monroe, NC) for GFP fluorescence

quantification (1 fish/well in 100 µL media for estrogenic effect quantification; 5

fish/well in 200 µL media for anti-estrogenic effect quantification). After 3 minutes

centrifugation at 3000 rpm, readings were taken using a fluorescence plate reader

(PerkinElmer 2030, Waltham, Massachusetts). Data was analyzed using Microsoft Excel.

Student t-test one-tail statistics were performed between treatment groups and DMSO

control group. Significance is presented at p ≤ 0.05(*) or p ≤ 0.01(**).

2.9 Chemical library screening

Tg(5xERE:GFP) embryos were arrayed into 96-well U-shape black plates and 96-well

black plates with clear bottom (Greiner Bio-One, Monroe, NC) in 100 µL of E3 solution

at 3 dpf (two embryos per well). The positive control for this assay was 10 µM E2, and

the negative control was 1% DMSO. Chemicals were added into the 96-well plates using

a VIAFLO eight-channel adjustable spacing electronic pipette (Integra, Hudson, NH).

For the primary screening, a 3mM working stock solution of each compound was

prepared from both ToxCastTM Phase I and NIH Clinical Collection I libraries, and 1 µL

of the working stock solution was added to each well of the 96-well plate, resulting in a

final screening concentration of 30 µM. 4-8 embryos (2-4 wells) were analyzed for each

compound. At 5 dpf, zebrafish were anesthetized with 0.002% eugenol. GFP fluorescence of the fish was imaged in the 96-well black plates with clear bottom

(Greiner Bio-One, Monroe, NC). GFP quantification of the fish in the 96-well U-shape black plate (Greiner Bio-One, Monroe, NC) was performed using the plate reader, as described above. Compounds in the ToxCast library that caused lethality of the embryos at 30 µM were rescreened at lower concentrations.

2.10 Bioinformatics

2.10.1 Biological function inference via pathway analysis and tissue enrichment

The corresponding human homologues to the differentially expressed zebrafish genes

(p≤0.01, fold change ≥|±1.4|) were identified using ZFIN (http://zfin.org/) and Ensembl

(http://www.ensembl.org). Gene ontology (GO) annotation biological processes enrichment of the estrogen responsive human homologues was performed by using

Pathway Studio (Ariadne, MD). Fisher’s exact test was used to calculate the p-value of each functional categories; p≤0.05 was considered significant.

Zebrafish estrogen responsive gene expression locations were categorized according to

ZFIN-Anatomy functional analysis at NIH-DAVID bioinformatics platform

(http://david.abcc.ncifcrf.gov/tools.jsp). Fisher’s exact test was used to calculate the P- value of the functional categories; p≤0.05 was considered significant.

2.10.2 Comparison of estrogen-regulated gene expression between embryos and adult male fish

Datasets of estrogen-induced gene expression in male adult zebrafish were obtained from the GEO database (Gene Expression Omnibus; http://www.ncbi.nlm.nih.gov/geo/) (GEO accession # GSE27707). Series matrix.txt files with the log2-normalized ratios for all samples were downloaded and One-way ANOVA was used for the statistical analysis of the treatment groups and control groups. Benjamini-Hochberg false discovery rate correction (FDR) was applied to the raw p-value. FDR q-value ≤0.01 and fold change

≥|±2.0 | were chosen as a cut off for the differentially expressed genes. To compare the estrogen target genes of embryos to those of adult fish, Venn-diagram analysis was performed.

2.10.3 Comparison of estrogenic pesticides among transgenic zebrafish assay and in vitro assays

Estrogenic pesticide datasets screened using in vitro assays are maintained at the ToxCast

Database (ToxCastDB) (http://actor.epa.gov/actor/faces/ToxCastDB/Home.jsp). The estrogen-related assays include Attagene Factorial trans ERα assay (ATG_ERα_TRANS),

Attagene Factorial cis ERE assay (ATG_ERE_CIS), NCGC ERα Agonist Assay 34

(NCGC_ERα_Agonist), NCGC ERα Antagonist assay (NCGC_ERα_Antagonist),

Novascreen Human ER (NVS_NR_hER), Novascreen Bovine ER (NVS_NR_bER) and

Novascreen Mouse ERa (NVS_NR_mERα) assays. Hits from these assays were obtained from http://actor.epa.gov/actor/faces/ToxCastDB/GenesAssocAssays.jsp. Venn diagrams were generated to illustrate the overlapping hits among the different assays.

Chapter 3. Estrogen Signaling during Zebrafish Development

3.1 Introduction

Zebrafish embryos have been used as a model for studying EEDs for decades (reviewed

in [111]), however, not much is known about estrogen signaling during embryonic

development. It is known however that the expression of different ER subtypes is tissue

specific and time dependent during zebrafish embryonic development [99,103-109].

Morpholino knock down of esr2a efficiently decreases the formation of neuromasts,

showing a direct role for esr2a in their development [109]. Exogenous E2 treatment at

concentrations up to 1 µM does not perturb normal embryonic development, but at 3 µM,

reduced hatching and pericardial edema have been observed, and at 10 µM or higher, fish

display a bent tail [103] and abnormal formation of the chondrocranium [130]. Treatment

of zebrafish during 48-168 hpf with an aromatase inhibitor, which induces estrogen

deficiency, causes neurobehavioral deficits, including decreased tactile response,

swimming movements, vestibular behavior, and pectoral fin and eye movements [131].

After prolonged treatment, the fish die by cardiac arrest. These effects can be blocked by

a simultaneous addition of estrogen [131], implicating the functional link to this pathway.

Estrogen deficiency also significantly diminishes thickness in most retinal layers,

suggesting that estrogen is important for normal eye development [132].

Although several biomarkers (such as Vitellogenins and Brain specific aromatase B) have

been established for assessing estrogenic effects of the EEDs in zebrafish, more E2-target

genes need to be identified in view of the complexity of estrogen signaling. Several

studies have been performed in adult zebrafish using a large-scale transcriptomic 36

approach. In one study, total mRNA from adult male zebrafish was analyzed on a custom-made microarray set containing about 16,000 oligonucleotide probes [133].

Approximately 1,000 estrogen-responsive genes were identified, including the already known target genes vtg1, vtg3 and esr1. Three other studies analyzed gene expression changes in liver of adult male zebrafish after E2 treatment using 14-16k microarray platforms [113,134,135], and identified hepatic E2-responsive genes. While the first study describes that the estrogen target genes are highly represented within functional categories of cell proliferation, apoptosis and gene expression, the other studies report that estrogen target genes are involved in metabolism. This observation logically reflects liver function [113,133-135]. Together, these studies identify a number of previously unknown estrogen target genes in the entire adult male organism or livers of zebrafish that potentially could serve as new biomarkers; however, a bioinformatic comparison of the genes described in the different publications has not yet been performed.

In our study, a whole-genome analysis of estrogen-regulated genes in zebrafish embryos at different developmental stages was performed. An Agilent zebrafish gene expression microarray with 44,000 probes was used to analyze organism-wide expression changes induced by E2, and bioinformatics was applied to interpret the consequences of these changes for early tissue development. The embryonic E2-target genes were further compared to previously published estrogen responsive genes in male adult zebrafish to identify potential biomarkers that can be used to detect xenoestrogenic exposures both to embryos and adult zebrafish.

3.2 Results

3.2.1 Distinct gene expression profiles are regulated by E2 during early zebrafish development

To increase the understanding of how estrogen acts on zebrafish early development, we performed a microarray analysis of whole-genome gene expression changes at different developmental time points. First, we made a dose response curve of E2-induced activation of the known estrogen targets vtg1 and esr1 by qPCR to determine which dose of E2 to use for the microarray experiments. Wild-type zebrafish embryos were treated with E2 at concentrations ranging from 0.01 nM to 1 μM from 3 hpf to 4 dpf with daily media exchange. We have previously shown that 1 μM E2 is the highest concentration that zebrafish embryos can tolerate without showing obvious phenotypic abnormalities

[103], and thus we did not investigate higher concentrations. Embryos were pooled and collected at 4 dpf for RT-qPCR. The expression of both vtg1 and esr1 was significantly induced by E2 treatment at 100 nM and maximally induced at 1 μM (Figure 3.1 A and B).

The 18S rRNA gene was used as a reference gene in the RT-qPCRs, since the expression of this gene was not influenced by E2 at the developmental stages or in the adult fish

(Figure 3.1 C).

We then performed the microarray analysis of whole-genome gene expression changes at different developmental time points. E2 treatment at 1 μM started at 3 hpf and embryos were pooled and collected at 1, 2, 3, and 4 dpf for RNA preparation. A total of 28 arrays, probed with cDNA prepared from 3 biological replicates for each control and E2-treated

fish at 1 and 2 dpf, and 4 biological replicates for 3 and 4 dpf were used for the

transcriptome analysis. Batch effects from the different biological replicates were removed using Partek Genomics Suite v 6.3 (http://www.partek.com/) and residual variance was analyzed by Principal Components Analysis (PCA) (Figure 3.2).

Transcriptome profiles identified a total of 298, 219, 1016, and 444 probes significantly altered by E2 treatment at 1, 2, 3, and 4 dpf, respectively (p≤0.01, absolute fold change

≥|±1.4|) (Table 3.1). Out of these, 136 genes were successfully annotated at 1 dpf; 104 genes at 2 dpf; 576 genes at 3 dpf; 204 genes at 4 dpf (Table 3.1). The annotated genes were further sorted by fold change. The top 15 up and down regulated genes upon E2 treatment for each time point are shown in Table 3.2-3.5. Comparing individual gene expression changes at each time point, distinct sets of genes were up or down-regulated by E2 at the different time points. In fact, Venn diagram analysis showed that only vtg1 expression was significantly activated by E2 treatment at all four time points (p≤0.01,

fold change ≥|±1.4|) (Figure 3.3A). The expression of 6 genes was co-regulated at 1 and

2 dpf, 9 genes at 1 and 3 dpf, 3 genes at 1 and 4 dpf, 29 genes at 2 and 3 dpf, 7 genes at 2 and 4 dpf, and 19 genes at 3 dpf and 4 dpf. Co-regulated genes among different time

points are shown in Table 3.6.

Then a Hierarchical clustering analysis of the significantly altered probes (p≤0.005) was

performed using the Pearson correlation algorithm. For this clustering we used a p-value

cutoff at p≤0.005 at which a clear developmental pattern of the differentially expressed

genes independently of E2 treatment was evident (Figure 3.4). The expression of a set of

Figure 3.1 Dose-response curves of vtg1 and esr1 expression in zebrafish embryos. Zebrafish embryos were treated with increasing doses of E2 continuously for 4 days and the mRNA expression levels were determined by RT-qPCR. (A) Relative vtg1 mRNA expression. (B) Relative esr1 mRNA expression. Asterisk denotes significant differences. (C) Reference gene 18S rRNA expression in the 1-4 dpf embryos and adult fish upon 1µM E2 treatment (relative to DMSO treatment). (**p<0.01; unpaired Student’s t-test compared to the controls; n=2 biological replicates; 3 technical replicates within each biological replicate). Abbreviations-vtg1: vitellogenin 1; esr1: estrogen receptor 1 and 18S rRNA: 18s ribosomal RNA.

Figure 3.2 Principle components analysis of microarray samples. (A) Untreated samples (vehicle only 0.1% DMSO). (B) Samples treated with E2.

Table 3.1 Numbers of differentially expressed genes at different developmental stages

p ≤ 0.005 p ≤ 0.01 p ≤ 0.01(Fold change ≥ |±1.4|) 1 dpf Total probes 140 323 298

Probes with gene 74 156 140 symbols Gene number 70 151 136

2 dpf Total probes 158 376 219

Probes with gene 114 249 116 symbols Gene number 101 230 104

3 dpf Total probes 418 1277 1016

Probes with gene 301 853 625 symbols Gene number 267 778 576

4 dpf Total probes 368 745 444

Probes with gene 228 461 227 symbols Gene number 208 425 204

genes was altered by E2 treatment at each time point (Figure 3.4). Table 3.1 details the

numbers of altered probes and genes at p-value cut-offs at 0.01 and 0.005. Selected genes

were validated with RT-qPCR (Figure 3.5). Details of oligonucleotide primer sequences

are shown in Table 3.7. Overall, there was a high concordance between the microarray

and the RT-qPCR results. For the up-regulated genes, qPCR confirmed the regulation of

vtg1, vtg3, vtg5, cyp19a1b, esr1, and f13a1a by E2, although at 1 time point (2 dpf) we only detected the up-regulation of vtg3 by RT-qPCR and not by microarray (Figure 3.5A).

Expression of esr1 was significantly up-regulated in RT-qPCR at 2, 3, and 4 dpf, but only at 4 dpf in microarray datasets (Figure 3.5A). The fold changes between the RT-qPCR and microarray results varied slightly when the induction was very high, but in general, the results showed high concordance.

For the down-regulated genes, the gene expression changes measured by RT-qPCR were

also consistent with results obtained by microarray analysis. Expression of hpx, agxtb,

fabp10a, fkbp5, klf9, pnp4b, zgc: 110053, nxf1, f2, and zgc: 92590 was significantly down-regulated based on both microarray and RT-qPCR (Figure 3.5B). The expression of the esr2a, esr2b, dlgap1a, and rbp2a genes was also confirmed by RT-qPCR, and the expression levels were not significantly changed at any of the four time points, which

was in concordance with the microarray data (Figure 3.5C).

3.2.2 Similar biological processes are regulated by E2 at different stages during

zebrafish development

In order to predict the consequences of gene-regulation, the human homologues to the

zebrafish estrogen responsive genes were identified through the ZFIN (http://zfin.org/)

and Ensembl zebrafish Zv9 databases (http://www.ensembl.org/Danio_rerio/), followed

by Gene Ontology (GO) enrichment of the human homologues using Pathway Studio

mammalian database. Approximately half of the annotated differentially regulated genes

were successfully identified with human homologues (64 (47.06%) at 1 dpf, 68 (65.38%)

at 2 dpf, 374 (64.93%) at 3 dpf, and 112 (54.90%) at 4 dpf). The human homologues to

the 15 most up- and down-regulated estrogen-responsive genes at different time points

are shown in Tables 3.2-3.5. A Venn diagram in Figure 3.3B shows human homologues

of differentially expressed genes co-regulated at the different time points and Table 3.6

details the co-regulated human homologues.

GO annotation analysis based on the human homologues revealed a significant

enrichment in several biological processes upon E2 exposure at the four time points

(Table 3.8). The broad functional categories included: metabolic process, transcription, transport, and signal transduction. Also, genes for phosphorylation, immune response and multicellular organism development categories were co-regulated at all four time points.

Apoptosis and cell proliferation categories were enriched from the differentially expressed genes at 3 and 4 dpf (Table 3.8). Under the broad functional categories, more specific sub-categories are shown in Table 3.9-3.13.

Table 3.2 Top 15 up- and down-regulated transcripts at 1 dpf upon E2 treatment (E2 vs control)

Gene Human p-value Fold-Change Genbank Accession Symbol homologue Up-regulated genes vtg3 1.23E-04 6.21 AF254638 dcaf13 DCAF13 7.60E-05 4.73 NM_200129 vtg1 8.65E-04 3.82 NM_001044897 taar19p TAAR6 2.42E-03 3.58 NM_001199914 col10a1 COL10A1 7.61E-03 3.39 NM_001083827 ptprc PTPRC 4.78E-03 2.71 CK127788 zgc:153505 DPYD 4.53E-03 2.70 NM_001077308 ppp1r1c PPP1R1C 7.26E-05 2.59 NM_001002538.1 dusp3 DUSP3 1.13E-03 2.50 NM_001044307 mettl16 METTL16 1.22E-02 2.43 BC131878 ugt5c3 4.79E-03 2.38 NM_001128714 rh50 6.42E-04 2.30 NM_131547 c1qtnf2 C1QTNF2 3.26E-03 2.22 XM_695699 hoxd12a HOXD12 5.56E-03 2.20 NM_001126486 igsf21b IGSF21 1.39E-03 2.18 NM_001110473 Down-regulated genes znf644 ZNF644 9.11E-03 -5.32 CT700944 nitr3d 1.21E-03 -4.43 NM_198355 ddc DDC 3.07E-03 -3.94 NM_213342 gria4a GRIA4 1.05E-02 -3.82 NM_214806 opn1lw2 OPN1LW 5.92E-04 -3.57 NM_001002443 prnp PRNP 1.16E-02 -3.49 NM_205586 zgc:193725 5.67E-03 -3.40 EH551598 npy7r 9.50E-03 -3.29 NM_001007218 slc7a11 SLC7A1 1.40E-02 -3.13 XM_001919391 zgc:112320 4.79E-03 -3.13 NM_001159826 nr1h5 NR1H5 8.09E-03 -3.05 NM_001123241 dpp6b DPP6 3.21E-03 -3.04 NM_001115122 hmgcll1 HMGCLL1 1.76E-03 -2.95 NM_001110400 sulf2l 3.38E-04 -2.71 NM_001003833 grid2 GRID2 8.27E-05 -2.59 NM_001004123

Table 3.3 Top 15 up- and down-regulated transcripts at 2 dpf upon E2 treatment (E2 vs control)

Human Gene Symbol p-value Fold-Change Genbank Accession homologue Up-regulated genes f13a1a F13A1 1.64E-09 28.39 NM_001076711 cyp19a1b CYP19A1 2.65E-04 7.58 NM_131642 vtg1 3.28E-05 4.50 NM_001044897.2 dnah10 DNAH 4.64E-03 4.02 XM_693415 vipr1 VIPR1 1.10E-03 3.68 NM_001013353 amh AMH 2.08E-05 2.88 NM_001007779 asb15b ASB15 1.49E-02 2.80 NM_001039890 desmb DES 6.35E-03 2.41 NM_001077452 coch COCH 2.54E-03 2.33 NM_001003823 or110-2 OR1Q1 1.05E-02 2.01 NM_001128411.1 map1s MAP1S 3.68E-03 1.91 XM_688813 mxtx1 DUXA 7.99E-03 1.73 NM_131560

rxraa RXRA 1.17E-03 1.68 NM_001161551.1 zgc:172270 1.35E-02 1.61 NM_001114568 nr0b2a NR0B2 4.58E-03 1.53 NM_001256191 Down-regulated genes fkbp5 FKBP5 3.40E-08 -15.81 NM_213149 klf9 KLF9 1.87E-05 -4.28 NM_001128729 pglyrp2 PGLYRP2 4.66E-03 -3.62 NM_001045166 fabp10a 5.83E-04 -3.37 NM_152960 zgc:162180 2.86E-03 -2.94 NM_001089446 pnp4b 2.96E-04 -2.90 NM_205643 hpx HPX 4.95E-04 -2.86 NM_001111147 ddc DDC 6.01E-04 -2.80 NM_213342 epd 5.48E-05 -2.71 NM_131005 kcnh6 KCNH6 1.16E-02 -2.61 XM_693502 rhcga RHCG 3.18E-03 -2.59 NM_001089577 unc13d UNC13C 8.50E-04 -2.57 CT598129 cxcr3.1 CXCR3 6.34E-03 -2.55 NM_001089430 nr1d2a 1.99E-03 -2.50 NM_001130592 ctsbb CTSB 8.66E-05 -2.40 NM_001110478

Table 3.4 Top 15 up- and down-regulated transcripts at 3 dpf upon E2 treatment (E2 vs control)

Human Gene Symbol p-value Fold-Change Genbank Accession homologue Up-regulated genes cyp19a1b CYP19A1 1.92E-08 35.10 NM_131642 vtg1 2.00E-08 7.91 NM_001044897 batf 1.07E-03 6.06 NM_001045059 c14orf135 C14ORF135 1.22E-02 5.70 CT698302 cd74b CD74 1.51E-03 5.60 NM_131372 cyp11a1 CYP11A1 3.39E-03 5.20 NM_152953 dazl DAZL 1.32E-02 4.97 NM_131524 hoxb9a HOXB9 1.32E-02 4.74 NM_131121 ccdc37 CCDC37 5.17E-03 4.51 NM_001076760 morc3b MORC3 1.12E-02 4.51 NM_001003579 dcst2 DCST2 7.94E-03 4.36 XM_679060 ubn2 UBN2 2.95E-03 4.12 NM_001104940 h1m H1FOO 7.88E-03 4.09 NM_183071 exosc6 EXOSC6 1.14E-02 3.98 BC065602 f13a1a F13A1 9.44E-05 3.88 NM_001076711 Down-regulated genes fabp10a 2.23E-07 -7.22 NM_152960 fabp6 FABP6 2.01E-04 -7.11 NM_001002076 fkbp5 FKBP5 5.42E-06 -5.44 NM_213149 zgc:193725 2.39E-04 -4.61 EH551598 hpx HPX 3.45E-04 -4.46 BC056563 dupd1 DUPD1 1.78E-03 -4.28 NM_001039837 itln3 ITLN2 1.53E-03 -4.25 NM_001159584 ca4b CA4 6.62E-03 -4.23 NM_001166211 serpinf2 SERPINF2 3.53E-05 -4.10 XM_683705 rhcga RHCG 1.66E-05 -4.00 NM_001089577 rhcgb RHCG 5.79E-03 -3.97 NM_207082 slc6a19b SLC6A19 6.03E-03 -3.70 NM_199736 tdo2a TDO2 1.25E-02 -3.69 NM_001102616 c3b C3 1.21E-02 -3.61 NM_131243 shbg SHBG 2.75E-04 -3.50 NM_001007151

Table 3.5 Top 15 up- and down-regulated transcripts at 4 dpf upon E2 treatment (E2 vs control)

Human Gene Symbol p-value Fold-Change Genbank Accession homologue Up-regulated genes vtg4 1.08E-15 1218.93 NM_001045294 vtg3 1.88E-14 646.00 AF254638 vtg1 3.12E-17 522.25 NM_001044897 cyp19a1b CYP19A1 2.28E-10 102.00 NM_131642 vtg2 1.67E-08 44.03 NM_001044913 f13a1a F13A1 3.35E-09 16.08 NM_001076711 vtg5 4.86E-15 14.96 NM_001025189 wipf2 WIPF2 8.35E-03 9.73 NM_001002165 lhbeta1 1.31E-03 5.05 NM_205623 zgc:153138 2.62E-03 4.72 NM_001076641 esr1 ESR1 1.94E-04 4.13 NM_152959 wdhd1 WDHD1 6.08E-04 3.60 NM_001002726 prop1 PROP1 4.55E-03 3.43 NM_001177461 amh AMH 1.40E-06 3.07 NM_001007779 tmem232 TMEM232 5.48E-04 2.71 NM_001100053 Down-regulated genes pfkfb4 PFKFB4 4.34E-04 -8.97 XM_002666441 upp2 UPP2 3.87E-04 -6.62 NM_200144 ahsg AHSG 7.52E-11 -5.61 NM_001100029 cbln8 3.88E-06 -5.56 NM_001110109 zgc:92590 4.68E-04 -4.76 NM_001007054 hpx HPX 3.40E-05 -4.32 NM_001111147 il11b IL11 1.35E-02 -3.68 XM_003200549 il34 IL34 1.17E-02 -3.63 NM_001082955 asb11 ASB11 8.89E-03 -3.37 NM_214792 zgc:171951 LGALS9 1.69E-03 -3.32 NM_001102630 casp8ap2 1.57E-04 -3.26 NM_001014334 gcga GCG 8.25E-03 -2.79 NM_001008595

armc4 ARMC4 1.26E-02 -2.67 XR_117794 mettl8 METTL8 9.81E-03 -2.54 NM_001007336 chrna2b CHRNA2 6.90E-03 -2.53 XM_692206

Figure 3.3 Distinct sets of genes are regulated by E2 during different times of zebrafish development. (A) Venn diagram illustrating the number of differentially expressed genes (p≤0.01, fold change ≥|±1.4|) that were regulated in common at the different time points. (B) Venn diagram of the human homologues of gene transcripts from (A).

point (1 at each time

eatment groups at different time

the biological replicates the biological

ssion levels between E2 and vehicle treated groups ssion levels between reflect up-regulated genes, blue cells reflect down- reflect up-regulated genes, blue cells on profiles of E2 and vehicle tr . ) The colored cells show the mean expression level of cells show the mean The colored <0.005 p Figure 3.4 Clustering of geneexpressi dpf and 2 dpf: n=3; 3 4 n=4). Red cells regulated genes and black cells reflect unchanged expre ( points.

Table 3.6 Co-regulated estrogen responsive genes at 1 dpf, 2 dpf, 3 dpf and 4 dpf upon E2 treatment

Overlapping Co-regulated genes Human homologues of the co- time points regulated genes 1, 2, 3 and 4 vtg1 dpf 1 and 2 dpf epd, c2cd2l, ddc, loc556849 DDC si:ch211-225b11.1, vtg1 1 and 3 dpf dnm1, lhcgr, loc100005283, DNM1 si:ch211-12e13.2, si:ch73- 138i16.1, zgc:193725, vtg3, epd, vtg1 1 and 4 dpf pomca, vtg3, vtg1 POMC

2 and 3 dpf ankrd9, anxa4, cr352285.1, ANKRD9, DNAJB5, FKBP5, cx28.9, dnajb5, fabp10a, fkbp5, HMGCS1, KLF3, KLF9, LCN15, hmgcs1, klf3, klf9, lcn15, NR1D2, PIK3R3, PNPLA3, loc559127, lpin1, nr1d2a, pik3r3, RHCG, SERPINF2, AMH, pnp4b, pnpla3, rhcga, serpinf2, CYP19A1, F13A1, HPX, si:dkey-30c15.2, wu:fc15e02, zgc:112242, amh, cyp19a1b, f13a1a, hpx, rgs2, vtg1, epd 2 and 4 dpf amh, cyp19a1b, f13a1a, hpx, AMH, CYP19A1, F13A1, HPX, rgs2, ctsbb, vtg1 CTSB, UNC13C 3 and 4 dpf bdp1, cbln8, dcakd, eif4e1b, C3, CA4, DCAKD, EIF4E1B, gcga, ins, lipc, loc559783, GCG, GREB1, KRT20, LIPC, sult1st3, vtg5, zcchc2, PFKFB4, RGS2, TRIM29, zgc:100868, amh, cyp19a1b, ZCCHC2, AMH, CYP19A1, f13a1a, hpx, rgs2, vtg1, vtg3 F13A1, HPX 1, 2 and 3 epd, vtg1 dpf 2, 3 and 4 amh, cyp19a1b, f13a1a, hpx, AMH, CYP19A1, F13A1, HPX dpf rgs2, vtg1 1, 3 and 4 vtg3, vtg1 dpf 1, 2 and 4 vtg1

Table 3.7 Primer sequences used for the RT-qPCR validation of estrogen responsive genes from the microarray

Genes Forward primer sequences (5’-3’) Reverse primer sequences (5’-3’) Genebank accession no. Reference gene (18s ribosomal RNA) 18s* (1) TCGCTAGTTGGCATCGTTTATG CGGAGGTTCGAAGACGATCA BX296557 Up-regulated genes esr1#(2) CAGGACCAGCCCGATTCC TTAGGGTACATGGGTGAGAGTTTG NM_152959 vtg1$(3) ACTACCAACTGGCTGCTTAC ACCATCGGCACAGATCTTC NM_170767 vtg3$(3) CAGATGGCTTTATCGGCGTGAC CACGGCAGGCCCATTGAAAC AF254638 vtg5%(4) CCAAAAATTGTCACCACTTATGCT CTTCATTCCTCCATGATATGCTTA BC142783.1 cyp19a1b AAAGAGTTACTAATAAAGATCCACC TCCACAAGCTTTCCCATTTCA AF226619 GGTAT f13a1a GCCCGTATTGCTTTTGCC GTGGGCCTGCTTGTTTTCTG NM_001076711.1 eif4e1b GACACGGTCGAGGACTTCTG GTGTGCCTATGTTGCTTGGC NM_131454.1 cyp11a1 GAGGCCTCAGGAATGTCCAC GGTCCACGCGTCTACATTGA NM_152953.2 dazl TGTTCGTCGGCGGTATTGAT TGACACTGACCGAGAACTTCG NM_131524.1 zp3 TCCAGCCAGTGGGTCTGACTA CCAACAATTGCACCATCAGTCC NW_003040435.2 cpn1 TGAGGGCTCTTTTTGCCGTC GACGTACTTGAACTCCGGCT NW_001877244.3 Down-regulated genes agxtb TCATCAGCTGGTTTGGTGCACCC GAGCAGCTCCCAATGATGCCAC NM_213162.1 pnp4b CGCATGCTGCACATCTTGGGGA TGAGCGAGAGACCCATGACCCG NM_205643.1 hpx CTCATAAAGGCAAACCTGGTG TGGACAGCTCAGCCTTGCCA NM_001111147 fabp10a AGCTGGTCTGCAGAACTGACCGA TGGTGGTTCCTCCGACTGTCAGC NM_152960.1 fkbp5 TCCACGAACCAGTGCCCGACT AGTGGACGAACACCCTGTCCC NM_213149.1 klf9 CACGGAAGCGCGACCGACTG GAAAGGGCCTCTCACCGGTGTG NM_001128729.1 zgc:110053 GCCAACTCGCCACAGCCAGAC CGAACTCCGGGCTCACACCGA NM_001020562.1 nxf1 GGCCTCGGAAGAGCAGAAATCTTC TCATTGAGGTGACCCAGTGCTGA XM_001343386.3 zgc:92590 GAGCCCAGTGTGAGGGTGCGTA CCGCCAGAATCTCCCTGACATGC NM_001007054.1 f2 ACTGTCAGGAGGGAGACCTG CGCTCTCCACAGTCTAGCTC NM_213390.1 sult1st3 TTCACTACTTCACTGACAACTGG CAGCTCTGTTCCGAATGGTATC NM_183348.2 Non-changed genes dlgap GAAGCTCCCGCCGCCAGTAC GCGCGTTTGGCTGCCATGAG XM_680713.4 rbp2a ACTGTTAAGACCCTGGTAAAGTGGG GAATCTCCAAGTGAAGCAAGTCTCC esr2b#(2) CGCTCGGCATGGACAAC CCCATGCGGTGGAGAGTAAT AAH86848 esr2a#(2) CTCACAGCACGGACCCTAAAC GGTTGTCCATCCTCCCGAAAC NM_001045184

References:

* (1) McCurley A and Callard G. (2008). Characterization of housekeeping genes in zebrafish: male-female differences and effects of tissue type, developmental stage and chemical treatment. BMC Molecular Biology. 9(102): 1471-2199

# (2) Chandrasekar G, Archer A, Gustafsson JA, Lendahl MA. (2010).Levels of 17b-Estradiol Receptors Expressed in Embryonic and Adult Zebrafish Following In Vivo Treatment of Natural or Synthetic Ligands. PLoS ONE. 5 (3): e9678.

$ (3) Meng X, Bartholomew C and Craft JA (2010). Differential expression of vitellogenin and oestrogen receptor genes in the liver of zebrafish, Danio Rerio. Anal Bioanal Chem: 396: 625-630

% (4) Sawyer S, Gerstner K and Callard G. (2006). Real-time PCR analysis of cytochrome P450 aromatase expression in zebrafish: Gene specific tissue distribution, sex differences, developmental programming, and estrogen regulation. General and Comparative Endocrinology 147: 108–117.

Figure 3.5 Comparison of E2 regulated genes analyzed by microarray or RT- qPCR. (A) Relative mRNA expression of the up-regulated genes vtg1, vtg3, vtg5, esr1, cyp19a1b and f13a1a at different time points as determined by RT-qPCR and microarray analysis. (B) Relative mRNA expression of the down-regulated genes hpx, fkbp5, fabp10a, agxtb, pnp4b, nlf1, klf9, zgc:92590, zgc:110053 and f2 at different time points as determined by RT-qPCR and microarray analysis. (C) Relative mRNA expression of the non-changed genes esr2a, esr2b, rbp2a and dlgap1a at different time points as determined by RT-qPCR and microarray analysis. White bars represent microarray results and black bars RT-qPCR results. Asterisk denotes significant difference (*P<0.05, **P<0.01; unpaired Student’s t-test compared to the controls), n≥3 biological replicates except for genes fabp10a, agxtb and zgc:110053 which were 2 biological replicates; each replicate consists of 30 pooled embryos. Abbreviations:vtg1: vitellogenin 1; vtg3: vitellogenin 3; vtg5: vitellogenin 5; esr1: estrogen receptor 1; esr2a: estrogen receptor 2a; esr2b: estrogen receptor 2b; cyp19a1b: cytochrome P450, family 19, subfamily A, polypeptide 1b; f13a1a: coagulation factor XIII, A1 polypeptide a; hpx: hemopexin; fkbp5: FK506 binding protein 5; fabp10a: fatty acid binding protein 10a; agxtb: alanine-glyoxylate aminotransferase b; pnp4b: purine nucleoside phosphorylase 4b; nxf1: nuclear RNA export factor 1; klf9: krueppel-like factor 9; f2: coagulation factor II ; rbp2a: retinol binding protein 2a and dlgap1a: discs, large (Drosophila) homolog-associated protein 1a.

Table 3.8 Gene ontology biological process functional group enrichment

Category* 1 dpf 2 dpf 3 dpf 4 dpf Percent Percent Percent Percent p-value p-value p-value p-value (%) (%) (%) (%) Metabolic 15.63 1.43E-02 28.13 3.02E-02 24.93 1.14E-17 38.32 2.53E-06 process Regulation of 10.94 1.24E-02 12.50 3.12E-02 10.03 3.93E-03 12.15 2.13E-02 transcription Transport 17.19 5.01E-04 23.44 2.04E-02 18.70 2.45E-13 28.04 9.82E-05 Signal 28.13 6.69E-05 17.19 4.02E-02 15.72 3.62E-04 18.69 6.83E-03 transduction Response to chemical -- -- 7.81 9.46E-05 3.52 1.12E-06 6.54 9.02E-06 stimulus Apoptosis 3.13 4.33E-01 6.25 6.52E-02 4.34 7.48E-03 5.61 3.51E-02 Cell -- -- 3.13 2.04E-01 2.98 5.37E-03 4.67 1.12E-02 proliferation Phosphoryla 7.81 1.56E-05 3.13 2.03E-04 5.42 1.07E-04 3.74 2.03E-02 tion Multicellular organismal 7.81 1.09E-02 4.69 1.93E-02 4.61 2.67E-02 3.74 4.86E-03 development Immune 4.69 2.00E-03 4.69 1.82E-02 2.17 1.76E-02 3.74 1.73E-02 response *Category represents main functional group, but p-value may represent subgroups of the main groups. Bold p-values represent statistically significant categories (p<0.05).

Table 3.9 GO terms sub-grouped into the metabolic process category (in italics).

Category 1 dpf 2 dpf 3 dpf 4 dpf Percent Percent Percent Percent p-value p-value p-value p-value (%) (%) (%) (%) Metabolic 15.63 1.43E-02 28.13 3.02E-02 24.93 1.14E-17 38.32 2.53E-06 process Hormone biosynthetic 3.13 9.91E-03 3.13 4.38E-02 2.71 9.28E-08 4.67 2.63E-04 process Cellular nitrogen compound 6.25 9.55E-04 1.56 6.21E-01 0.54 1.01E-01 4.67 2.52E-02 metabolic process RNA metabolic 3.13 5.42E-02 3.13 2.02E-01 1.36 1.23E-01 5.61 5.42E-04 process ATP catabolic 1.56 4.56E-01 3.13 3.89E-01 1.63 7.40E-04 5.61 2.81E-02 process

Lipid metabolic -- -- 7.81 2.94E-02 5.15 2.78E-06 9.35 7.83E-04 process

Carbohydrate metabolic -- -- 3.13 5.09E-01 4.07 3.79E-04 10.28 1.80E-04 process

Glucose metabolic -- -- 1.56 7.13E-02 1.63 1.39E-02 4.67 4.24E-03 process

Cellular protein metabolic -- -- 1.56 7.13E-02 2.17 7.58E-02 4.67 9.10E-02 process

Xenobiotic metabolic -- -- 1.56 5.06E-01 2.71 8.26E-05 4.67 7.17E-03 process

Energy reserve metabolic -- -- 1.56 4.05E-01 2.44 3.97E-05 4.67 1.96E-03 process

Bold p-values represent statistically significant categories (p<0.05).

Table 3.10 GO terms sub-grouped into the transport category (in italics)

Category* 1 dpf 2 dpf 3 dpf 4 dpf Percent Percent Percent Percent p-value p-value p-value p-value (%) (%) (%) (%) Transport 17.19 5.01E-04 23.44 2.04E-02 18.70 2.45E-13 28.04 9.82E-05 Transmem brane 4.69 2.49E-01 14.06 1.52E-02 11.38 3.37E-12 14.02 2.57E-03 transport Ion 10.94 5.82E-04 6.25 3.98E-01 7.32 4.41E-06 8.41 9.82E-02 transport

Sodium ion 1.56 2.75E-01 1.56 4.08E-02 3.52 4.53E-07 1.87 3.42E-01 transport Calcium ion 1.56 2.68E-01 1.56 4.08E-02 1.90 6.57E-03 0.93 2.34E-02 transport Amino acid 3.13 1.78E-03 3.13 4.06E-02 1.63 5.09E-04 1.87 1.03E-01 transport

Carbohydra ------1.08 8.97E-03 1.87 1.23E-01 te transport Lipid -- 3.13 2.16E-01 0.81 2.16E-01 2.80 3.26E-02 -- transport Bold p-values represent statistically significant categories (p<0.05).

Table 3.11 GO terms sub-grouped into the signaling pathways category (in italics)

Category* 1 dpf 2 dpf 3 dpf 4 dpf Percent Percent Percent Percent p-value p-value p-value p-value (%) (%) (%) (%) Signal 28.13 6.69E-05 17.19 4.02E-02 15.72 3.62E-04 18.69 6.83E-03 transduction Synaptic 9.38 9.39E-05 4.69 4.09E-02 2.44 2.12E-02 1.87 3.45E-01 transmission Steroid hormone mediated -- -- 4.69 1.72E-04 1.08 8.42E-03 2.80 9.33E-03 signaling pathway Cell-cell -- -- 4.69 1.41E-01 2.17 5.45E-02 4.67 1.90E-03 signaling Intracellular protein 1.56 2.30E-01 1.56 4.40E-01 1.63 1.95E-03 3.74 1.73E-02 kinase cascade Activation of MAPK 3.13 1.51E-02 -- -- 1.36 3.70E-03 1.87 3.84E-02 activity Cell surface receptor linked 4.69 1.29E-02 1.56 7.00E-01 1.36 8.51E-04 0.93 8.72E-01 signaling pathway Wnt receptor 1.56 4.68E-02 3.13 4.27E-02 0.81 4.43E-01 4.67 1.08E-02 signaling pathway Toll-like receptor 1 3.13 9.65E-03 4.69 6.75E-03 1.36 8.51E-04 -- -- signaling pathway Glutamate signaling 4.69 3.50E-06 ------pathway Nerve growth factor 1.56 3.74E-01 3.13 2.77E-01 2.98 2.39E-05 2.80 2.59E-01 receptor signaling pathway Bold p-values represent statistically significant categories (p<0.05).

Table 3.12 GO terms sub-grouped into the multicellular organismal development category (in italics)

Category* 1 dpf 2 dpf 3 dpf 4 dpf Percent Percent Percent Percent p-value p-value p-value p-value (%) (%) (%) (%) Multicellular organismal 7.81 1.09E-02 4.69 1.93E-02 4.61 2.67E-02 3.74 4.86E-03 development Embryo 4.69 1.09E-02 4.69 1.03E-01 1.90 6.60E-02 1.87 8.84E-02 development Fat cell differentiatio -- -- 4.69 1.91E-04 1.08 2.87E-03 1.87 1.38E-02 n Brain 1.56 4.01E-01 4.69 1.22E-02 1.08 1.77E-03 3.74 1.83E-05 development

Liver 1.56 2.03E-01 3.13 1.93E-02 1.08 2.67E-02 2.80 4.86E-03 development Muscle -- -- 3.13 3.95E-02 1.08 3.17E-02 2.80 1.27E-03 development Gland 1.56 4.08E-02 1.56 4.08E-02 0.81 1.77E-02 3.74 1.70E-04 development

Blood vessel -- -- 3.13 3.54E-02 0.54 2.39E-01 1.87 1.71E-02 development

Kidney -- -- 1.56 4.05E-01 1.08 2.91E-02 0.93 5.88E-01 development Bold p-values represent statistically significant categories (p<0.05).

Table 3.13 GO terms sub-grouped into the response to chemical stimulus category (in italics)

Category* 1 dpf 2 dpf 3 dpf 4 dpf Percent Percent Percent Percent p-value p-value p-value p-value (%) (%) (%) (%) Response to chemical -- -- 7.81 9.46E-05 3.52 1.12E-06 6.54 9.02E-06 stimulus Response to -- -- 4.69 2.62E-02 2.44 1.59E-04 6.54 9.60E-05 hormone Response to estrogen 1.56 1.92E-01 1.56 2.18E-01 2.44 2.59E-04 4.67 8.79E-05 stimulus Response to 4.69 7.99E-02 9.38 2.51E-02 6.50 2.36E-07 9.35 2.85E-06 drug Bold p-values represent statistically significant categories (p<0.05).

Table 3.14 Number of genes enriched by the NIH-DAVID tissue enrichment platform

Time point 1 dpf 2 dpf 3 dpf 4 dpf No. % No. % No. % No. % Input genes 135 100 103 100 575 100 203 100 NIH David identified 100 74.07 68 71.92 397 69.04 144 70.94 genes Enriched genes in ZFIN anatomy 43 43 38 55.88 178 44.84 64 44.44 category

Table 3.15 ZFIN anatomy functional chart of E2 responsive genes enriched by NIH DAVID analysis tool

Term Count % p-value Genes 1 dpf gria2a, gria3a, oprd1a, grk7a, ppp1r1c, ptprn2, angptl1, lrrc4c, Brain 9 9.18 3.20E-02 scn4ba, gria4a, epd Retinal photoreceptor 3 3.06 1.20E-02 grk7a, zgc:112320, ntm layer Cephalic 3 3.06 5.80E-03 smyhc2, myhb, hspb8 musculature 2 dpf Brain 9 13.2 1.50E-02 bsk146, rgs2, nr1d2a, hmgcs1, vipr1, tcf7l2, rxraa, epd, cyp19a1b Liver 8 11.8 1.70E-02 pnp4b, rgs2, hpx, pglyrp2, hmgcs1, vtg1, cyp19a1b, fabp10a Pancreas 3 4.41 2.30E-03 hsd11b2, spon1b, atp1a3a primordium Intestinal bulb 4 5.88 8.80E-03 bcmo1, pnp4b, rgs2, zgc:110176 YSL 6 8.82 1.40E-02 nr0b2a, mxtx1, hmgcs1, gnsb, arl5c, nfkbiaa Ovary 4 5.88 1.90E-02 amh, vtg1, vipr1, cyp19a1b Testis 5 7.35 9.70E-04 amh, vtg1, vipr1, fabp10a, cyp19a1b 3 dpf Ventral 5 1.26 1.40E-03 fgf19, cadm4, nr4a1, etv1, cyp19a1b telencephalon cyp1b1, mre11a, lhcgr, hmgcs1, kmo, si:ch211-93f2.1, il17rd, scn1ba, atp2b1b, cyp19a1b , fabp10a, ttr, sult2st1, myd88, Liver 40 10.1 3.00E-08 zgc:103559, vtg3, vtg1, vtg5, unc45a, shbg, cebpd, hkdc1, fa2h, atp7a, sall4, pnp4b, rgs2, uox, hpx, f2, ghrl, ripk2, srd5a2a, zgc:92111, zgc:153921, eaf2, nr5a5, acad11, lipc, fabp6, c3b cpa4, ins, ctrb1, neu3.3, ghrl, zgc:92111, gcga, zgc:66382, try, Pancreas 10 2.53 6.80E-04 ela3l sult2st1, dab2, pnp4b, myd88, rgs2, neu3.3, si:ch211-93f2.1, Intestinal bulb 10 2.53 5.20E-03 srd5a2a, nr5a5, acad11 cyp1b1, slc2a11l, bcl2, slc13a1, slc26a6l, ripk2, eaf2, atp2b1b, Kidney 10 2.53 9.30E-03 illr4, fabp6 hkdc1, ms4a17a.5, fa2h, zgc:101040, sypl2a, atp1a3b, kmo, Pronephric duct 16 4.04 5.70E-02 slc20a1a, cep70, tnfrsf1a, dab2, sall4, myd88, slc13a1, ahcyl2, ip6k2 cyp1b1, zp3, lhcgr, atp2b1b, cyp19a1b, amh, bcl2, vtg3, ghrl, vtg1, Ovary 12 3.03 3.30E-03 nr5a5, vtg5, fabp6 amh, cyp1b1, lhcgr, ghrl, vtg1, nr5a5, atp2b1b, vtg5, cyp19a1b, Testis 9 2.27 1.10E-02 fabp10a 4 dpf pgr, rarab, ahr1a, irak3, bdnf, rgs2, rnaset2, eif4a2, lhbeta1, cdk5, Brain 12 8.45 3.20E-02 disc1, cyp19a1b slc43a1a, zgc:174260, lhbeta1, esr1, zgc:113054, ahsg, cpn1, Liver 19 13.4 1.70E-06 cyp19a1b ,bdnf, serpina7, rgs2, atic, hpx, rnaset2, vtg3, vtg4, vtg1, vtg2, rpia, lipc, vtg5 Pancreas 6 4.23 6.00E-04 zgc:92590, gip, ins, rwdd3, gcga, amy2a Neuromast 4 2.82 9.80E-03 bdnf, vtg3, zgc:56382, sb:cb252 Ovary 8 5.63 3.10E-04 amh, rnaset2, lhbeta1, vtg3, esr1, vtg1, vtg2, vtg5, cyp19a1b Testis 6 4.23 1.90E-03 pgr, amh, rnaset2, lhbeta1, vtg1, vtg5, cyp19a1b 61

Most of the sub-categories also overlap at the four developmental stages. Hormone

biosynthetic process, steroid signaling pathway, and response to estrogen stimulus

validated the estrogenic effects during zebrafish development upon E2 treatment. To

summarize, although distinct sets of genes were regulated by E2 at different time points,

the biological processes that these genes affect were similar.

3.2.3 Tissue-specific enrichment of E2 responsive genes in embryonic zebrafish

To infer the expression location of the E2 responsive genes from the microarray experiment, we performed knowledge-based data mining on the tissue enrichment by using ZFIN_Anatomy functional analysis at NIH DAVID bioinformatics microarray analysis platform (http://david.abcc.ncifcrf.gov/tools.jsp). Enriched categories represent the common expression locations of the clusters of E2 responsive genes, suggesting that these tissues are estrogen-responsive tissues. The numbers of genes identified by NIH

DAVID and enriched from ZFIN_Anatomy functional analysis are listed in Table 3.14.

E2-responsive genes were significantly enriched in the brain at all four time points although at 3 dpf, ventral telencephalon, which is a subgroup of the category “brain”, was enriched (Table 3.15). Liver, pancreas, and reproductive organ categories emerged at 2, 3, and 4 dpf. At 1 dpf, an enriched category was retinal photoreceptor, while intestinal bulb showed up at 2 and 3 dpf. E2-responsive genes were enriched in the kidney and pronephric duct categories at 3dpf, and in the neuromast category at 4 dpf (Table 3.15).

To visualize in which tissues E2 activates gene expression, transgenic zebrafish embryos expressing GFP driven by 5×ERE were used [125]. The fish were treated with 1 μM E2

62 continuously from 3 hpf to 6 dpf and were observed daily for GFP expression under a fluorescence microscope. In the absence of E2 treatment, fluorescence signal was detected before 1 dpf. To determine whether this signal was caused by maternal load of

GFP or zygotic transcription/translation of it, we crossed Tg(5xERE:GFP) fish with wild- type DZ fish. The embryos from female Tg(5xERE:GFP) fish crossed with male wild- type DZ fish showed a similar fluorescence signal to the homozygous Tg(5xERE:GFP) embryos (Figure 3.6A-D). Conversely, when crossing male Tg(5xERE:GFP) fish with female DZ fish, the embryos did not show any fluorescence (Figure 3.6E-H). Thus, we concluded that the initial fluorescence of the Tg(5xERE:GFP) embryos represents maternal load of GFP expression. At 1 dpf, the maternal GFP fluorescence had faded and the zygotic GFP expression appeared mainly in the head region after E2 treatment but not in untreated embryos (Figure 3.7A-C and results not shown). In the presence of E2, strong GFP fluorescence was detected from 2 dpf in the presumptive liver progenitor cells and persisted as the liver developed (Figure 3.7D-U). GFP-positive cells were also detected in the brain and heart valves at 4 dpf, but the expression was very weak. At 5 and 6 dpf, the GFP expression was more visible in the brain, pre-optic nerves, ear region, heart valves, liver, and pancreas (Figure 3.7J-U). The GFP expression at the later time points was in accordance with the previous report on these transgenic fish [125].

To conclude, according to the target tissue-enrichment analysis, endogenous E2- responsive genes are mainly expressed in the liver, pancreas, and at various locations of the brain during early zebrafish development (Table 3.15). This is consistent with the regions of GFP expression in Tg(5×ERE:GFP) transgenic fish after E2 induction

(Figure 3.7). The results are also in accordance to the biological pathway analysis

showing that the most regulated process is the one of metabolism, presumably taking

place in the liver. We thus conclude from these findings that liver, pancreas and brain are

sensitive organs for estrogen treatment, or exposure to other estrogenic compounds,

during early zebrafish development.

3.2.4 Comparison of estrogen signaling between adult and embryonic zebrafish

Another study on the analysis of estrogen target genes in the whole organism of zebrafish

reported microarray datasets from adult male zebrafish (GEO accession # GSE27707)

[133]. To investigate whether there were any E2 target genes regulated in common between adult zebrafish and embryos in response to E2, we compared our results to the report of Lam and colleagues [133]. Our significantly up-regulated and down-regulated genes (p≤0.01, fold change ≥|±1.4|) were compared with estrogen responsive genes in

male adult fish (q≤0.01, fold change ≥|±2.0|) using the same fold induction as in the

Lam et al. publication [133], by Venn diagram analysis (Figure 3.8A and B). The

expression of vtg1 was commonly up-regulated in male adult fish and all four embryonic

stages and vtg3 was up-regulated in all the stages except 2 dpf. Expression of cyp11a1,

eif4e1b, dazl and zp3 was up-regulated in 3 dpf embryos and adult males, while

expression of esr1, atic and cpn1 was up-regulated in 4 dpf embryos and adult males

(Figure 3.8A). The down-regulated genes in common in both embryos and adult males

were sult1st3, f2 and zgc:56382 (Figure 3.8B). The expression of selected co-regulated

genes was validated in embryos and adult fish by RT-qPCR (Figure 3.9A and B). Details

of primer sequences are shown in Table 3.7. The known estrogenic markers vtg1, vtg3 64

and esr1 showed strong induction by E2 in the adult fish (Figure 3.9A) as did the

expression of cpn1 (Figure 3.9B). Although f13a1b and cyp19a1b were not identified as

E2 target genes in the adult fish microarray dataset, we found that their expression was also highly induced by E2 in the male adult fish (Figure 3.9A). The expressions of eif4e1b, cyp11a1, and zp3 were all up-regulated in both 3 dpf embryos and adult males upon E2 treatment as predicted, but to much lower levels. Expression of dazl was up-

regulated in the 3 dpf embryos but not in the adult males. Expression of sult1st3 was

significantly down-regulated in the 3 dpf embryos but not 4 dpf embryos and adult males.

Expression of f2 was down-regulated in both 3 dpf embryos and adult males (Figure

3.9B). Since the expression of vtg1, vtg3, cyp19a1b, esr1, cpn1, and f13a1a was up-

regulated more than 20 times, they can likely serve as biomarkers for estrogenic signaling.

GFP expression has been detected in the liver and pituitary organ in the Tg(5xERE:GFP)

adult fish upon treatment with the synthetic estrogenic compound DES by Daniel

Gorelick and colleagues [125], which is similar to our results for tissue enrichment and

GFP expression patterns in zebrafish embryos (Table 3.15 and Figure 3.7). Furthermore,

we analyzed the GFP expression in E2 treated adult transgenic reporter fish. GFP expression was detected in different regions in the heart of female and male fish. In addition, GFP expression was found in the intestine and gall bladders in both male and female fish, however, it was stronger in the female than in male fish. Ovary from the female and testis from the male fish all had strong GFP expression (Figure 3.10). In summary, the estrogen signaling between the developmental zebrafish embryos and adult fish is distinct, although some estrogen responsive tissues and genes are in common.

3.3 Discussion

3.3.1 E2 regulates estrogen responsive genes in a stage-dependent manner

We identified E2 responsive genes that were regulated at 1-4 dpf during zebrafish

development through transcriptome analysis. We further showed E2 activated

transcription of a distinct set of genes in a stage- and tissue-specific manner during

zebrafish development.

From hierarchical clustering and Venn diagram analysis, we found that most of the

differentially expressed genes upon E2 treatment were characteristic for different stages

during zebrafish development (Figure 3.3 and 3.4). At 1 dpf, among the top up- and

down-regulated genes were genes that encode proteins involved in ubiquitination

(dcaf13), phosphorylation (ptprc, ppp1r1c, and dusp3), development of tissues (hoxd12a, znf644, and opn1lw2), immune system (igsf21b, nitr3d), and synapse function (ddc, gria4a, and grid2) (Table 3.2). At 2 dpf, the most highly activated (28 fold) gene transcript was f13a1a, which encodes coagulation factor VIII, and the most down- regulated (15 fold) gene transcript was fkbp5, an immunophilin involved in protein folding and trafficking (Table 3.3). At 3 dpf, the cyp19a1b gene showed the most highly activated transcription following E2 treatment (35 fold) (Table 3.4). Other activated transcripts were transcription factors (batf, hoxb9a, ccdc37, and morc3b) and ubiquitin- related gene transcripts (ubn2 and exosc6). The top down-regulated transcripts at 3 dpf were two fatty acid binding proteins (fabp10 and fabp6), and other down-regulated genes encoded proteins involved in transport, including the heme transporter hpx, the

ammonium transporters rhcga and rhcgb, and slc6a, which transport amino acids in the kidney. Finally, the most up-regulated transcripts at 4 dpf were clearly the vtgs4, 3, 1, 2,

and 5 (1218, 646, 522, 44 and 15 fold up-regulated by E2, respectively) (Table 3.5). The

most highly down-regulated gene transcripts at this stage were pfkfb4 (9 fold), a hypoxic

induced kinase/phosphatase, and upp2 (6 fold), which is a uridine phosphatase.

Although most E2 responsive genes were different at each stages of zebrafish development, a few genes were co-regulated at several time points. The well-known estrogenic biomarkers vtg1, vtg3, vtg5, esr1, and cyp19a1b were found to be E2-target genes in our study, confirming that our results are in line with previously published results [136-139]. Expression of vtg1 and vtg3 was not only up-regulated by E2 at embryonic stages but also in male adult fish, as published previously (Figure 3.5 and

Figure 3.9, and [133]), verifying the status of these genes as reliable estrogenic biomarkers.

In addition to the well-known estrogenic biomarkers, some further transcripts regulated by E2 at several developmental stages were found to be potential biomarkers for estrogen signaling. Transcription of ependymin (epd), encoding a brain extracellular glycoprotein involved in memory and neuronal regeneration (further discussed below), was down- regulated by E2 at 3 time points. Transcription of anti-Mullerian hormone (amh), encoding a steroidogenic enzyme, was induced at 3 time points. In previous studies, amh was reported to be induced in both prepubertal and adult rats upon methoxychlor exposure, a previously commonly used pesticide, now banned in the US because of its estrogenic disrupting properties properties [140,141]. The transcripts of coagulation 67

Figure 3.6 Maternal transfer of GFP expression in Tg(5xERE:GFP) transgenic fish at 5 hpf in the absence of E2. (A, B) Tg(5xERE:GFP) transgenic fish embryos. (C, D) Embryos from cross of female Tg(5xERE:GFP) transgenic fish and male wild type DZ fish. (E, F) Embryos from cross of male Tg(5xERE:GFP) transgenic fish and female wild type DZ fish. (G, H) Wild type DZ fish embryos. A, C, E and G, bright-field images; B, D, F and H corresponding GFP fluorescence images; Scale bars, 500 μm.

Figure 3.7 E2-induced GFP expression in Tg(5xERE:GFP) transgenic fish during development. Zebrafish larvae were treated with 1 μM E2 (in 0.1% DMSO) from 3 hpf and imaged at 1 dpf (A-C), 2 dpf (D-F), 3 dpf (G-I), 4 dpf (J-L), 5 dpf (M-R) and 6 dpf (S-U). A, D, G, J, M, P and S, bright-field images; B, E, H, K, N, Q and T corresponding GFP fluorescence images; C, F, I, L, O, R and U, overlay of bright- field and GFP images. A-C and S-U, lateral view; D-O, dorsal view; P-R, ventral view; anterior to the left. Scale bars, 100 μm. 69

Figure 3.8 Venn diagrams showing overlapping estrogen responsive genes among embryos and male adult fish. (A) Overlapping up-regulated genes changed by E2 in male adult fish versus 1-4 dpf embryonic fish. (B) Overlapping down-regulated genes changed by E2 in male adult fish versus 1- 4 dpf embryonic fish.

Figure 3.9 Validation of co-regulated differentially expressed genes in embryos and adult males using RT-qPCR. (A) Relative mRNA expression of the up-regulated genes vtg1, vtg3, esr1, f13a1a and cyp19a1b in adult males upon E2 treatment. (B) Relative mRNA expression of the genes eif4e1b, cyp11a1, dazl, zp3, cpn1, sult1st3 and f2 in 3 dpf and 4 dpf larvae as well as in adult males upon E2 treatment. Abbreviations-eif4e1b: eukaryotic translation initiation factor 4e 1b, cyp11a1 (cytochrome P450, subfamily XIA, polypeptide 1), dazl: deleted in azoospermia-like, zp3: zona pellucida glycoprotein 3, cpn1: carboxypeptidase N, polypeptide 1, sult1st3: sulfotransferase family 1, cytosolic sulfotransferase 3.

A B C D

E F G H

I J K L

Figure 3.10 GFP expression in the heart, intestine, gall bladder, ovary and testis of E2 treated adult Tg(5xERE:GFP) fish. A-D, heart; E and F, intestine; G and H intestine and gall bladder; I and J, ovary, K and L, testis. A, B, E, F, I, and J, organs from female fish; C, D, G, H, K, and L, organs from male fish; A, E, I, C, G, and K, bright-field images; B, F, J, D, H, and L, corresponding fluorescence images. Scale bar, 1 mm.

factor XIII, A1 polypeptide a (f13a1a), regulator of G-protein signaling 2 (rgs2), and

hemopexin (hpx) were highly altered by E2, as described above. HPX protein levels are decreased in estrogen-treated menopausal women, which is consistent with the decrease of hpx transcripts upon E2 treatment in our study [142]. Genes that were regulated at two

embryonic time points include ankyrin repeat domain 9 (ankrd9), FK506 binding protein

5 (fkbp5), fatty acid binding protein 10a (fabp10a), purine nucleoside phosphorylase 4b

(pnp4b), proopiomelanocortin a (pomca), and dopa decarboxylase (ddc) (Table 3.2-3.5).

Further investigations will be required to understand the role of estrogen regulation of

these genes during development.

Comparison of the E2 target genes that were changed in early developmental stages with

the ones changed in adult fish suggested that a few genes were co-regulated in embryonic

and adult fish, e.g. f13a1a, eukaryotic translation initiation factor 4e 1b (eif4e1b), zona

pellucida glycoprotein 3 (zp3), and carboxypeptidase N, polypeptide 1 (cpn1). In

summary, several previously known and new E2 target genes were found to be regulated

in a time-dependent manner. Out of these, a number of genes, which were regulated at

several time points, are potential estrogenic biomarkers for estrogen exposure.

3.3.2 E2 regulates estrogen-responsive genes in a tissue-specific manner

Putative estrogen target tissues, in which the E2-responsive genes have been reported to

be expressed, were predicted by using ZFIN anatomy functional analysis at NIH DAVID

bioinformatics platform (Table 3.14). Differentially expressed genes were enriched in

brain, liver, and pancreas at 2-4 dpf, which is in accordance with the GFP-expressing

73 tissues in the Tg(5xERE:GFP) transgenic fish (Figure 3.7 and [125]). Similar estrogen responsive tissues have also been reported by another group using a similar 3×ERE-

Gal4ff/UAS-GFP double transgenic fish [143]. However, the latter fish model showed additional GFP expression in the muscle fibers in the somites at 4 dpf, a finding that was not made with the Tg(5xERE:GFP) fish [125,143]. Supporting a role for E2 in muscle, one of the GO categories enriched at 2-4 dpf was “muscle development” (Table 3.12).

In our analysis, the tissue with the most E2 responsive genes, such as fabp10a, pnp4b, hpx, rgs2, and the vtgs, was the liver (Table 3.15). Liver is the major organ for metabolism, detoxification, and homeostasis. In the Tg(5xERE:GFP) transgenic fish, liver expression of GFP can be detected at 35 hpf, which is in agreement with the in situ hybridization of hdac1 of developing liver [83]. Consistent with our results, the fish liver is a main target for both endogenous and exogenous estrogens, and the classical estrogen biomarkers in fish include the liver expressed genes vtgs and esr1.

Another tissue in which E2 responsive genes were enriched was the pancreas. At 2 dpf, a cluster of gene transcripts were enriched in the pancreas primordium. At 3 and 4 dpf, some gene transcripts were enriched in the pancreas. Similar to the mammalian pancreas, the zebrafish pancreas includes both an exocrine/duct compartment and endocrine part comprising alpha, beta, delta, epsilon, and pancreatic polypeptide-producing cells [84].

The endocrine pancreas is one of the major organs of zebrafish endocrine system secreting insulin, glucagon, PP, ghrelin, and somatostatin [84]. Our results showed that transcription of insulin (ins), ghrelin/obestatin preprohormone (ghrl1), and glucagon a

(gcga) were all down-regulated in the pancreas upon E2 treatment (Table 3.15). The 74

identification of pancreas specific E2-responsive genes is in agreement with the GFP

expression in the Tg(5xERE:GFP) transgenic fish, which is evident in the pancreas

(Figure 3.7). This suggests that estrogen receptors are present in the pancreas and that estrogen signaling plays a role in zebrafish pancreas development. Although no report has been published linking estrogens to pancreas function in zebrafish, the evidence for this connection in mammals is extensive. In particular the role of estrogens in regulation of proliferation, differentiation, and survival of β cells and of insulin synthesis and release has been described (reviewed in [85]).

Estrogen receptors, in particular ERβ (ESR2), are important for normal brain and

behavior development in rodents. Whereas ERα (ESR1) is the predominant ER in the

hypothalamus, controlling reproductive cycles, ERβ is expressed in the cerebral cortex,

the hippocampus, the cerebellum and the dorsal raphe (reviewed in [144]). In zebrafish,

estrogenic regulation of genes occurs at several locations, including the preoptic area,

olfactory bulb, and hypothalamus in the brain of the transgenic Tg(5xERE:GFP) fish

([125] and Figure 3.7). In accordance with the pattern of ERE-induced GFP expression,

the brain was one of the tissues that had many enriched gene transcripts following E2

treatment. At the earliest time point studied, several subunits of the α-amino-3-hydroxy-5-

methyl-4-isoxazoleproprionate (AMPA) type glutamate receptors showed up as E2 target

genes. These receptors mediate the majority of fast synaptic glutamate transmissions,

which promote neuronal growth, retraction, and elongation of glial processes,

proliferation and differentiation of retinal progenitors, and proliferation of cortical

progenitors. Specifically, gria2a, gria3a, and gria4a glutamate receptor subunits were

regulated by E2. The AMPA receptors are regulated by estrogen in rats and mice

[145,146]. The epd transcript, encoding a protein involved in neuronal plasticity,

neurobehavior (memory and aggression), and cold adaptation, and the neurobehavioral

disc1 transcript were also regulated by E2 treatment in embryonic zebrafish brain. The

disc1 gene has been associated with risk of schizophrenia, bipolar affective disorder and

major depression, the first two in a gender specific way [147]. Also, cyp19a1b (aromatase

B) was found to be regulated by estrogen in brain at 2-4 dpf. Although in the tissue

clustering, this gene was designated to several tissues, including liver [148], cyp19a1b

has been reported to be brain specific in zebrafish while cyp19a1a encodes aromatase in

the ovary [149]. Consistently, we did not detect any mRNA of cyp19a1b in adult fish

liver extracts by RT-qPCR (results not shown). More specifically, expression of

cyp19a1b is up-regulated by E2 in zebrafish radial glial cells [150].

A group of estrogen-responsive genes was shown to be expressed in kidney or pronephric

duct, including the cytochrome cyp1b1, the ATPases atp2b1b and atp1a3b (Ca++ and

Na+/K+ transporting, respectively), fatty acid-binding protein 6 (fabp6), and the solute carrier family members slc2a11l, slc13a1, slc26a6l, and slc20a1a. Cyp1b1 has been detected in the developing kidney during early murine development [151]. It metabolizes estradiol and plays an important role in normal embryonic development [152]. Renal dysfunction and inflammation associated with angiotensin II-induced hypertension of the mouse model is cyp1b1 dependent [153]. In line with the E2 regulation in the kidney

and/or pronephric duct by the four transcripts that belong to genes of the solute carrier family, slc2a11l (glucose transporter), slc13a1 (sodium/sulphate transporter), slc26a6l

(anion transporter), and slc20a1a (phosphate transporter), one of the biological functional groups that was regulated by E2 was “Transport”, as discussed below.

Besides liver, brain, pancreas, and kidney, the tissue enrichment in Table 3.15 also includes cephalic musculature and retinal photoreceptors at 1 dpf, and intestinal bulb, testis, and ovary at 2-4 dpf; however in the 5×ERE-transgenic fish we failed to detect

GFP expression in any of these organs at the early developmental stages. Intestine, testis, and ovary expression of GFP was, however, detected in the adult Tg(5xERE:GFP) transgenic fish (Figure 3.10). Normal morphogenesis of ovary and testis does not initiate until 10 dpf [154,155], but some genes controlling sex differentiation like amh might be altered by E2 treatment or play other roles at earlier stages of zebrafish development. In addition, Tg(5xERE:GFP) transgenic fish showed weak GFP expression in the heart valves from 4 dpf as well as in the adult fish (Figure 3.7 and 3.10), which is in agreement with observations from the two other groups on estrogen reporter GFP transgenic fish

[125,143]. However, we did not obtain any enrichment of estrogen responsive gene transcripts in the heart (Table 3.15). Finally, the tissue analysis showed that estrogen regulates genes in neuromasts, cells that depend on esr2a for their development [109].

Expression of brain-derived neurotrophic factor (bdnf), which is involved in the both the

development and maintenance of neuromasts [156], was up-regulated by E2. Although

this gene has not been shown to be regulated by estrogen in zebrafish before, many

reports have described an E2 induction of bdnf expression in mammals.

In conclusion, estrogen responsive gene transcripts were found to be expressed in various

tissues according to knowledge-based data mining; many of these tissues are in 77 concordance with the ones that have been identified by transgenic estrogen reporter fish, but some of them are novel E2 target tissues for zebrafish.

3.3.3 E2 regulates genes with similar biological functions during zebrafish development

Although the estrogen-responsive genes during early zebrafish development were expressed in a time-dependent and tissue-specific manner, the biological functional processes were similar across all time points. Analysis by GO-term biological processes enrichment for E2-regulated genes identified metabolic processes, regulation of transcription, transport, signal transduction, phosphorylation, development, and immune response to be significantly enriched at most time points (Table 3.8). Some subcategories of these main categories also overlapped among all the time points (Table 3.9-3.13).

However, some categories were not significantly enriched across all time points; apoptosis and cell proliferation categories were enriched only at 3 and 4 dpf, and the response to chemical stimulus category at 2-4 dpf.

In line with the liver being one of the major tissue targets for estrogenic signaling, E2 target genes were enriched in the category metabolism. The genes in this category encode both metabolic enzymes in the steroid and hormone pathways, as well as enzymes involved in lipid, nucleic acid, carbohydrate, protein, xenobiotic, and energy reserve metabolism (Table 3.9). The transcripts of the major estrogen-metabolizing genes

CYP19A1, encoding aromatase B (zebrafish homologue cyp19a1b), sulfotransferase

SULT1E1 (zebrafish homologue sult2st3), and hydroxysteroid (17β) dehydrogenase

HSD17B (zebrafish homologue hsd17b) were all differentially expressed in our study; expression of cyp19a1b and sult2st3 was up-regulated while hsd17b was down-regulated by E2. For lipid metabolism, lipc (encoding hepatic lipase) was down-regulated and apolipoprotein A-I (encoded by apoa), which is the major protein component of high density lipoprotein (HDL) in plasma, was up-regulated by E2. Studies of humans and primates confirm the regulation of LIPC and APOA-1 by E2. First, E2 is known to repress the transcription expression of LIPC in humans, which plays an important role in lowering the plasma level of HDL [157,158], a function requiring ESR1 [157]. Further, hepatic apoa1 is induced by E2 in the human hepatoma cell line HepG2 [159].

Furthermore, APOA-1 levels and production rate increase during postmenopausal estrogen replacement therapy [160]. Finally, metabolic studies on ovariectomized and hysterectomized baboons show that E2-treatment increases APOA-1 content of HDL

[161].

Another GO biological category that was enriched in our analysis was transport, including ion transport (specifically sodium and calcium ion transport), lipid transport, protein transport, and carbohydrate transport (including glucose transport) (Table 3.8 and

3.10). As described above, a group of solute carrier family genes was regulated by E2, including slc2a1 (glucose transporter), slc8a1 (sodium/calcium exchanger), slc5a5

(sodium iodide symporter), and slc34a2 (sodium phosphate transporter). Expression of slc2a1 is stimulated by E2 to increase glucose uptake in various tissues [162-164] and slc34a2 mRNA expression is increased by 50% in rat intestine after E2 treatment [165].

Expression of slc8a1 was up-regulated by E2 in rabbit hearts [166] and slc5a5 was up-

regulated by E2 in mammary gland [167,168] and breast [169,170]. However, the

expressions of these genes were all up-regulated by estrogen in these studies, but down

regulated in our study, suggesting that solute carrier family might play a different role

during zebrafish development than in mammals.

E2 treatment also regulated gene transcripts that were enriched in the signal transduction

and transcription factor categories (Table 3.8 and 3.11), which is in line with how E2 functions at a molecular level. E2 signaling pathways are mediated by estrogen receptors,

ESR1 and ESR2 (or zebrafish receptors esr1, esr2a and esr2b), as well as the membrane- bound estrogen receptor (GPER). Signaling through these receptors has been shown to cross talk with other receptors or transcription factors. The subcategories that were enriched in our GO analysis included steroid hormone mediated signaling pathway, G- protein coupled receptor protein signaling pathway, cell surface receptor linked signaling pathway, as well as MAPK pathway (Table 3.11), validating the genomic and non- genomic effects of E2 during zebrafish development. Expression of several receptors were differentially regulated by E2 treatment, including growth hormone releasing hormone receptor (ghrhr), retinoid X receptor α (rxra), esr1 itself, progesterone receptor (pgr), and nuclear receptor subfamily 0b2a, (nr0b2, SHP), which were all up- regulated by E2, and aryl hydrocarbon receptor (ahr), retinoic acid receptor α (rara), nuclear receptor subfamily 5a2 (nr5a2, LRH), luteinizing hormone/choriogonadotropin receptor (lhcgr), glucagon a (gcg), and androgen receptor(ar), which were down- regulated. It has been extensively reported that E2 induces pgr expression, which is in line with our data. Both esr1 and esr2 bind to the pgr promoter and esr1 induces pgr

80 expression. However, esr2 has been reported to reduce pgr expression [171] and an increase of ESR2:ESR1 ratio may suppress PGR expression and contribute to progesterone resistance [172]. AHR signaling is known to cross talk with ER signaling.

E2 has been reported to repress AHR trans-repression through binding to esr1 [173].

Expression of RARα was previously shown to be up-regulated by E2 in various tissues

[174-176], but our data showing the opposite suggests new roles for E2 on regulation of

RARα expression during zebrafish development. E2 also induces expression of SHP in the mouse and rat liver and in human HepG2 cells [177]. Given that LRH-1 is involved in estrogen production [178], the down-regulation seen in our data may suggest a negative feedback mechanism of E2 to LRH expression. Overall, our data show that induction of

E2 signaling translates to a crosstalk with several other receptors during zebrafish development.

E2 also changed the expression of genes involved in cell proliferation and apoptosis, such as cysteine-rich angiogenic inducer 61 (cyr61, CCN1), B-cell lymphoma 2 (bcl2), and caspase 3 (casp3). Expression of cyr61 has been reported to be induced by estrogen in breast cancer cells [179,180] and human myometrial explants [180], which is consistent with the up-regulation of cyr61 expression in our study. Expression of bcl2 was down- regulated in our study, but previous reports have shown both induction and repression of bcl-2 expression by estrogen, potentially in a tissue-specific manner [181-185].

Expression of casp3, a death protease activated during apoptosis, was also down- regulated during zebrafish development upon E2 treatment. E2 at neuroprotective doses blocks casp3 activation in the hippocampal CA1 of male gerbils [186]. In fetal

neuroepithelial cells, E2 strongly inhibits the activation of casp3 [186]. In accordance with the previous studies, expression of cyclin-dependent kinase 5 regulatory subunit 1a

(cdk5, p35) [187] and protein tyrosine phosphatase receptor type C (ptprc, CD45) [188]

was up-regulated upon E2 treatment.

Our data identified several phosphatase genes as being regulated by E2. Such genes

include dual specificity phosphatase 3 (dusp3), which dephosphorylates and inactivates various MAPKs like ERK and JNK [189,190], and protein phosphatase 1, regulatory

(inhibitor) subunit 1C (ppp1r1c). Finally, the category immune response was also enriched amongst E2-regulated gene transcripts (Table 3.8). In our gene list, interferon- gamma (ifng1-1), a vital immunoregulatory cytokine, was up-regulated, which is in agreement with previous studies showing an increase of IFNG secretion by estrogen treatment in mice and rats [191,192]. ESR1 increases IFNG expression [193,194] and

GPER can enhance IFNG production [195]. Tumor necrosis factor receptor superfamily

1a (tnfrsf1) was down-regulated in zebrafish embryos after E2 treatment, which is consistent with a study showing that E2 inhibits TNFR1 expression in breast adipose fibroblasts [186].

3.3.4 E2-activated blood coagulation pathways in zebrafish

In the GO analysis, the blood coagulation pathway, which is a subcategory of “response to wounding”, was enriched from the estrogen-responsive genes in 2-4 dpf embryos

(Table 3.13). In this pathway, expression of two genes f2 (human homologue F2) and f13a1a (human homologue F13A1), encoding coagulation factor II (thrombin) and factor

XIII, respectively, was verified by RT-qPCR in both E2 treated embryonic and adult

zebrafish (Figure 3.5 and 3.9).

To explore the link between E2 and blood coagulation, Pathway Studio was used to build a postulated pathway connecting estrogens and the differentially expressed genes enriched in the blood coagulation pathway (Figure 3.11). As seen in Figure 3.11, estrogens can regulate blood coagulation by directly regulating F2 (thrombin, zebrafish homologue f2), or by regulating F2 through either ERβ or ERα and CD40, a member of

the tumor necrosis factor (TNF) receptor superfamily. It has been reported that hormone

replacement therapy increased thrombin generation and fibrinolysis in postmenopausal women [3,79,81]. Oral contraceptive pills also increased thrombin generation and

increased plasmin generation in healthy women [196]. Estrogen administration enhanced

thrombin generation in rats, and increased thrombin in their uteri [197,198]. However,

thrombin was down-regulated in our dataset, suggesting a different role of E2 on thrombin expression during zebrafish development. The finding might also reflect the complex pathways mediating estrogen effects on blood coagulation. For example, it has

been reported that ERβ is membrane-associated in platelets and mediates a non-genomic

effect of E2 in human platelets through Src Kinase [199]. Furthermore, raloxifene, a

SERM, has been reported to reduce E2 induced expression of CD40 in follicular B cells

of ovariectomized mice [200]. High level of CD40 in the plasma of lung cancer patients

correlates with platelet and/or coagulation activation [201]. However, E2 has also been

reported to block the induction of CD40 protein expression on endothelial cells, mediated

by ERα [202]. These findings infer CD40-mediated estrogen effects on blood coagulation.

F13A1 (blood coagulation factor VIII, encoded by F13A1, zebrafish homologue f13a1a),

was also enriched in the blood coagulation pathway. Expression of f13a1a was highly up-

regulated by E2 in both embryonic and adult zebrafish (Figure 3.5 and 3.9), suggesting

that it can serve as a potential new estrogenic biomarker. F2 and F13A1 are reported to

be two biomarkers of deep venous thrombosis (reviewed in [203]). Oral contraceptive pills, estrogen replacement therapy, and SERMs are all associated with an increased risk of thrombosis [204,205]. Moreover, estrogens can markedly affect the coagulation system, with increased levels of procoagulant factors VII, IX, X, XII, and XIII, and reduced concentrations of the anticoagulant factors PS and AT. These effects on the coagulation system could be observed during pregnancy, treatment with oral contraceptives (OC) or hormone replacement therapy (HRT) use [206]. Tamoxifen, another SERM, prevents breast cancer but causes a risk of deep vein thrombosis [207].

To sum up, E2 activates blood coagulation pathways during zebrafish development.

Putative pathway building (Figure 3.11) suggested f2 and f13a1a might play important

roles in this activation.

In conclusion, our data reveal distinct differences in the cohort of E2 responsive genes

across developmental stages in the zebrafish. Although the differentially expressed gene

sets altered by E2 are different at the four embryonic developmental stages, the major GO

biological processes are in common. E2 activates target genes in a tissue specific manner

as visualized by GFP expression of Tg(5xERE:GFP) transgenic fish. Brain, liver, heart,

and pancreas are major estrogen responsive organs in developing embryos and adults.

However, Tg(5xERE:GFP) fish cannot mirror all E2-responsive transcriptional activation

and target tissues since the GFP expression is only ERE driven. Genes that are activated by an ERE-independent mechanism, such as the Gper-activated genes, will not be visible

with the Tg(5xERE:GFP) transgenic fish. Our study revealed E2-responsive genes independently of whether E2 targeted Esr1, 2a or 2b or Gper through genomic or non- genomic actions. The new target genes, such as f13a1a, cpn1, and zp3, are potentially important for estrogen signaling during development and might be used as biomarkers to score for endocrine disruption. To further the knowledge of how estrogen regulates embryonic development, or the impact of perturbed estrogen signaling by exposure to estrogen disrupting compounds, it will be necessary to map estrogen receptor type specific responses through selective estrogen modulators and knockdown/knockout of

Esr1, 2a, 2b, or Gper.

Figure 3.11 Postulated pathway of E2 regulation of blood coagulation during zebrafish development. The pathway was built using Pathways Studio 7 software. Only direct relationships are shown. Green oval represents E2 treatment. The proteins labeled with a red cloud represent significantly up- regulated expression after E2 treatment in the zebrafish embryos. The proteins labeled with a blue cloud represent significantly down-regulated expression after

E2 treatment. Red ovals represent proteins. Red diamonds represent ligands; Red ovals with docking sites represent transcription factors; Red cone represents receptors.

Chapter 4. Assessing Estrogenic and Anti-estrogenic Effects of

Ten Bisphenols using Zebrafish Larvae

4.1 Introduction

Bisphenol A (BPA), a well-known endocrine disrupting compound (EDC), has been

widely used to make epoxy resin and polycarbonate plastics since the 1950s. It is found

in products such as baby bottles, lining of beverage and food cans, dental materials,

medical equipment, computers, and cell phone casings. Diet, drinking water, indoor air,

indoor dust, and wastewater are all important sources of human exposure to BPA

(reviewed in [208]). It was recently described that another route of exposure is through

dermal exposure of BPA used in thermal paper [209]. In humans, BPA has been detected

in various tissues, urine, blood, and other body fluids (reviewed in [64]). Association of urinary (total) BPA and heart disease, diabetes, and liver toxicity has been made from adult human studies ([210], reviewed in [64]). BPA has been extensively reported to be

an estrogenic EDC acting through ERs causing many deleterious effects such as

abnormal fetal development [211], alteration of sex differentiation [212], reproduction

[213], brain function, and mood [214,215], as well as some types of cancers [216-218] in

animals. Moreover, babies and young children under six years old are estimated to be more sensitive to the effects of BPA than older children and adults [219,220]. Therefore,

BPA was banned in baby bottles in many places all over the world such as Canada, the

European Union, and the United States.

Concurrent with the increasing awareness of health risks linked to BPA, manufacturers of

the plastic industry have started to search for alternatives. These alternatives include the

bisphenol derivatives BPB, BPC, BPE, BPF, BPS, BADGE, BPAF, and BPAP. BADGE, one of the BPA-derivatives, has been commonly used in protective coatings inside food

and drink cans. However, just like BPA, BADGE has been found to migrate from the

coatings to the food (reviewed in [221]). BPF, which is also commonly used for

manufacturing of plastics, coatings, and resins, has been found in surface water, sewage,

sediments [222], soft drinks [223], and indoor dust [224]. BPS and BPB have been

detected in canned foodstuffs [225-227], and BPS and BPAF have been found in indoor

dust [224]. The other bisphenols, BPAP, BPC, BPCl, and BPE are used as raw materials

in the industry for producing polycarbonate, polyester, and epoxy resins, but research on

exposure levels and health effects of these compounds is very limited.

Reporter cell lines have been used for assessing estrogenicity of potential xenoestrogens.

In the MCF-7 breast cancer cell line transiently transfected with hERα and hERβ

expressing plasmids, BPA causes luciferase reporter gene activation [228]. Kitamura and

colleagues profiled the endocrine-disrupting activities of BPA and 19 of its derivatives

using an ERE-luciferase reporter assay in MCF-7 cell line [229]. They reported that BPA

and some of its derivatives, such as BPB, BPF, and BPAF, have estrogenic effects. In

addition, BPA, BPB, and BPF have weak anti-estrogenic effects [229]. Furthermore,

Matsushima and colleagues transiently co-transfected HeLa cells with ERα or ERβ

expressing plasmid as well as a plasmid containing 3xERE upstream of luciferase

reporter gene. Using these cell lines, they found that BPAF is a full agonist for ERα but

88 an antagonist for ERβ [230]. BADGE has no estrogenic effects in stably transfected

MCF-7 cells with the ERE-luciferase reporter gene [231,232]. Recently, Delfosse and colleagues showed that BPA, BPAF, and BPC-Cl act as partial agonists to the ERs showing both estrogenic effects and anti-estrogenic effects using stable HELN cell lines expressing ERs and a luciferase reporter [233].

Zebrafish (Danio rerio) serves as a complimentary, but important, model to cell-based models for studying bisphenols. The three zebrafish ERs (Esr1, Esr2a, and Esr2b) are all expressed from embryonic stages to adulthood, but with different organ distributions

[99,103]. Cosnefroy and colleagues reported that in stable ERE-luciferase zebrafish liver cell lines, BPA has high affinity for all three zfERs, but its estrogenic potency is much lower than that of E2, similar to what has been described for BPA and E2 using cells with human ERs [234]. Using the transgenic reporter zebrafish Tg(5xERE:GFP), Gorelick and

Halpern reported that 10 µM BPA activates GFP expression in the liver and heart valves of the transgenic fish, but 1 µM BPA only activates GFP expression in the heart valves

[125]. Lee and colleagues generated a similar transgenic fish line Tg(3×ERE:Gal4ff,

UAS:GFP) containing 3xERE and a Gal4ff-UAS system as well as GFP reporter. Using this fish line, they found that 5 µM BPA activates GFP expression in the heart, cranial muscle, otic vesicles, and retinal ganglions of the reporter fish [126].

Moreover, liver-expressed Vitellogenin1 (encoded by vtg1) and brain-expressed

Aromatase B (encoded by cyp19a1b) have been widely used as estrogenic biomarkers in teleost fish to assess environmental estrogens. BPA induces vtg1 expression in adult male zebrafish [113,235], developing larvae [138], and GFP expression in vtg1-GFP 89 transgenic reporter zebrafish [123]. Brion and colleagues reported that BPA can also induce GFP expression in cyp19a1b-GFP transgenic zebrafish [124]. However, according to Wang et al., while the expression of cyp19a1b is strongly up-regulated by synthetic estrogen 17-alpha-ethinylestradiol (EE2), it is down-regulated by BPA in minnow

Gobiocypris rarus juveniles [236]. Therefore, more estrogenic biomarkers need to be defined to assess xenoestrogens.

Although estrogenic effects of BPA have been assessed using zebrafish, the estrogenicity of bisphenols other than BPA has not. In this chapter, estrogenic and anti-estrogenic activities of 10 bisphenols were profiled and compared using different zebrafish assays.

First, the activation potential of the bisphenols on Tg(5xERE:GFP) transgenic reporter zebrafish was investigated, and GFP fluorescence was monitored and quantified.

Secondly, expression of several estrogen responsive genes from the microarray gene list discussed in chapter 3, including vtg1, vtg3, cyp19a1b, f13a1a, and esr1, was monitored in wild-type zebrafish after exposure to the bisphenols. Lastly, the anti-estrogenic activities of 10 bisphenols were investigated by co-treatment with EE2 in transgenic reporter fish embryos as well as wild-type fish. The GFP expression and mRNA expression of vtg1 and vtg3 was analyzed.

4.2 Results

4.2.1 Bisphenols activate GFP-expression in Tg(5xERE:GFP) transgenic zebrafish in a tissue-specific and dose-dependent manner

To assess the estrogenic effects of the ten bisphenols in vivo in zebrafish larvae,

transgenic zebrafish Tg (5xERE:GFP) were used to visualize the GFP expression upon

bisphenol exposure. A 4 dpf – 6 dpf exposure window was chosen because the larvae

showed low toxic effect of the chemical exposure and the positive GFP-fluorescence

intensity was high. The estrogenic activation of GFP-expression in selective tissues in

this transgenic fish has been described by Gorelick and Halpern [125] and in chapter 3.

Dose-response experiments ranging from 10 nM to 10 µM were performed for all 10

bisphenols (Table 4.1). EE2 was used as positive control, as it is more stable in zebrafish than E2 [237], and vehicle only (0.1 % DMSO) was used as negative control. The highest exposure dose for each bisphenol was chosen to be below the dose causing general toxicity to zebrafish larvae. GFP-expression was observed at 6 dpf using fluorescence microscopy and its intensity was quantified using a fluorescence plate reader. EE2 treatment induced GFP expression in the liver, pancreas, eyes, pre-optic area (POA), brain, and cells in the ear region (Figure 4.1, Table 4.2). Similarly to EE2, BPA, BPE,

BPF, and BPB exposure also activated GFP expression in the liver, pancreas, eyes, POA, and brain at 10 µM and 3 µM, but only BPA and BPF exposure activated GFP expression in the ear region (Figure 4.2). At 1 µM, these four bisphenols only activated GFP expression in the heart valves and brain. The expression pattern of the reporter fish exposed to 1 µM BPF is shown in Figure 4.3. No GFP expression was observed after 91

exposure to doses lower than 100 nM for BPA, BPE, BPB, and BPF. BPAF exposure

induced a similar GFP expression pattern as EE2 at 3 µM and 1 µM, but only activated

GFP expression in the heart valves and brain at 300 nM and 100 nM (Figure 4.2 and

Table 4.2). No GFP expression was observed after exposure to BPAF at doses lower than

10 nM (Table 4.2). The finding that GFP-specific expression is only detectable in the heart valves when exposed to low concentrations of bisphenols is in line with the previous report on BPA exposure of these fish [125].

Exposure to BPC, BPC(Cl), and BPAP activated GFP-expression in a different manner

compared to BPA, BPE, BPF, BPB, and BPAF. BPC exposure activated GFP-expression

in the heart valves, brain, pancreas, and the ear region, but only very weakly in the liver

at 3 µM and 1 µM (Figure 4.4). It only activated GFP-expression in the ear region and

brain at 300 nM and 100 nM. Exposure to BPC(Cl) at 3 µM and 1 µM activated GFP-

expression in the ear region, brain, and very weakly in the liver and heart valves, which is

different from the GFP-expression pattern of exposure to all of the other bisphenols

(Figure 4.4). Exposure at 300 nM and 100 nM of BPC(Cl) only activated GFP-expression in the ear region and brain (Figure 4.4). Exposure to BPAP only activated GFP- expression in the brain and ear region at 3 µM, but not at all at lower doses (Figure 4.4).

Exposure to BPS and BADGE did not activate any GFP-expression in the transgenic fish larvae at any doses tested (Table 4.2).

Table 4.1 Bisphenols for the study

Name CAS No. Structural formula

Bisphenol A 80-05-7

Bisphenol AP 1571-75-1

Bisphenol AF 1478-61-1

Bisphenol B 77-40-7

Bisphenol C 79-97-0

Bisphenol C (Cl) 14868-03-2

Bisphenol E 8-5-2081

Bisphenol F 87139-40-0

Bisphenol S 80-09-1

Bisphenol A 1675-54-3 diglycidyl ether

Table 4.2 GFP fluorescence expression patterns in Tg(5xERE:GFP) fish after bisphenol treatments

Chemicals Dosage (log [M]) -8 -7 -7.5 -6 -6.5 -5 EE2 All* All* All* All* -- -- Liver (faint), Liver, heart, brain, eyes, Heart BPA N# N Heart (faint) heart, brain, ears, ears, pancreas, and (faint) and pancreas kidney Liver, heart, Heart Liver, heart, Heart and brain, eyes, BPAF N and brain, eyes, ears, Toxic (heart edema) brain ears, and brain and pancreas pancreas Liver, heart, Heart and Heart and Liver, heart, brain, eyes, BPE N N brain, and brain brain and pancreas pancreas Heart Liver, Heart, brain, Heart and Liver (faint), BPF N N (some, eyes, ears, pancreas, brain heart, and brain faint) and kidney Liver (faint), Liver (some strong and Heart Heart and BPB N N heart, brain, and some faint), heart, (faint) brain pancreas brain, and pancreas Brain Liver, heart, Liver (faint), Brain and BPC N and brain, and heart, brain, ears, Toxic (heart edema) ears ears ears and pancreas Brain Liver (faint), Liver (faint), heart Brain and BPC(Cl) N and brain, and (faint), brain, and Toxic (heart edema) ears ears ears ears Heart (some, BPAP N N N N faint), brain, and Toxic (heart edema) ears BADGE N N N N N Toxic (heart edema) BPS N N N N N N Note: All* means the GFP expression was observed in the liver, heart, brain, eyes, ears, pancreas, and kidney. N# means no positive GFP expression was observed in the transgenic reporter zebrafish.

Figure 4.1 EE2 induced GFP fluorescence expression pattern in Tg(5xERE:GFP) fish larvae. 4 dpf Tg(5xERE:GFP) fish larvae were treated with EE2 (3 nM) for 48 hours. Pictures were captured at 6 dpf. A and B lateral view, anterior to the left; C dorsal view, anterior to the left; D, E and F, ventral view, anterior to the left. Scale bars, 200 μm (A, B, D and E) and 100 μm (C and F). Abbreviations: h, heart; p, pancreas; l, liver; of, olfactory bulb; hc, hair cells; POA, preoptic area; b, brain.

Lateral view Ventral view

BF GFP Overlay BF GFP Overlay

Figure 4.2 Comparison of the GFP expression patterns in Tg(5xERE:GFP) fish after exposure to EE2, BPA, BPAF, BPE, BPF and BPB. 4 dpf Tg(5xERE:GFP) fish larvae were treated with 3 nM EE2 (A), 10 µM BPA (B), 3 µM BPAF (C), 10 µM BPE (D), 10 µM BPF (E), or 10 µM BPB (F) for 48 hours. Pictures were captured with the same imaging parameters at 6 dpf. Scale bar, 200 μm.

BF GFP Overlay

Figure 4.3 Comparison of the GFP expression patterns between 10 µM and 1 µM BPF treatment. 4 dpf Tg(5xERE:GFP) fish larvae were treated with 10 µM BPF (A and B), or 1 µM BPF (B and D) for 48 hours. Images were captured at 6 dpf. GFP fluorescence intensity was adjusted to show the expression pattern. A and C, lateral view, anterior to the left; B and D ventral view, anterior to the left. Scale bar, 200 μm.

To conclude, BPA, BPAF, BPE, BPF, and BPB exposure induced a similar GFP-

expression pattern as EE2, with a strong GFP-expression in the liver. BPC, BPC(Cl), and

BPAP did not induce a strong GFP-expression in the liver, but induced expression in the

heart valves or ear region.

To quantify the GFP-expression of the transgenic fish larvae, we transferred the 6 dpf

transgenic fish larvae into 96-well U-shaped micro-plates and measured fluorescence in a

plate reader (1 fish/well). BPE induced the highest level of GFP fluorescence among all of the bisphenols (Figure 4.5). BPAF, BPA, BPF, and BPB all had significant activation of GFP expression at their highest concentrations compared to vehicle only, but the GFP intensity was lower than the EE2-induced intensity, indicative of their partial agonistic effects. These results are consistent with the observation of GFP expression from the fluorescence microscope. However, the relatively weak GFP expression in the heart valves, ear region, and brain induced by all the bisphenols except BADGE and BPS was not detected by the plate reader, which thus was a quantitative, but a less sensitive detection method than fluorescence microscopy. In conclusion, exposure to all the bisphenols, except BADGE and BPS, activated GFP-expression in Tg(5xERE:GFP)

transgenic zebrafish larvae in a dose-dependent and tissue-specific manner.

Lateral view Ventral view

BF GFP Overlay BF GFP Overlay

BPC 3 µM

BPC (Cl) 3 µM

BPAP 3 µM

Figure 4.4 Comparison of the GFP-expression patterns in Tg(5xERE:GFP) fish after exposure to BPC, BPC(Cl) and BPAP. 4 dpf Tg(5xERE:GFP) fish larvae were treated with BPC (3 µM), BPC(Cl) (3 µM) and BPAP (3 µM) respectively for 48 hours. Pictures were captured with the same parameters at 6 dpf. Arrows show GFP-positive tissues. Scale bar, 200 μm.

(Cl)

Figure 4.5 Bisphenol-induced GFP fluorescence quantification. 4 dpf Tg(5xERE:GFP) fish larvae were treated with EE2 (3 nM), BPA (10 µM), BPE (10 µM), BPF (10 µM), BPB (10 µM), BPS (10 µM), BPAF (3 µM), BPC (3 µM), BPC(Cl) (3 µM), BPAP (3 µM), BADGE (3 µM), or 0.1% DMSO for 48 hours. Media was renewed at 5 dpf. Larvae were transferred to 96-well plates for GFP quantification (1 fish/well). Results are expressed as relative fluorescence intensity above control (means ± SEM). Asterisk denotes significant difference (**p<0.01, unpaired student’s t-test compared to the DMSO control). n=12 biological replicates. This experiment was repeated 3 times.

100

4.2.2 EE2 and bisphenols induce expression of estrogen-responsive genes in zebrafish larvae

To investigate further the estrogenic effects of the bisphenols and to validate estrogenic biomarkers, mRNA expression levels of several estrogen-responsive genes from our microarray gene list were tested after bisphenol exposures. Expression of the well-known estrogenic biomarkers including liver-expressed markers vtg1 and vtg3, brain-expressed cyp19a1b, and the more ubiquitously expressed esr1 were first analyzed. The expression of the potentially new biomarker suggested in chapter 3, f13a1a, encoding blood coagulation factor XIII, was also investigated following bisphenol exposures. Treatment of 4 dpf zebrafish larvae with the positive control EE2, followed by RT-PCR showed that the mRNA expression of all five genes was up-regulated in a dose-dependent manner, reaching the activation plateau at around 300 nM (Figure 4.6 A), confirming that these genes are valid estrogenic biomarkers. Furthermore, ICI 182,780, an estrogenic antagonist, abolished the induction of vtg1, f13a1a, cyp19a1b, and esr1 expression by

EE2, confirming that these genes are downstream genes of ERs (Figure 4.6 B).

Exposure to the bisphenols induced different expression patterns of these genes in the zebrafish embryos. BPA, BPAF, BPE, BPF, and BPB exposure induced the expression of vtg1, vtg3, cyp19a1b, f13a1a, and esr1 in a dose-dependent manner, but to a lower level than EE2 did, confirming their partial agonistic effects (Figure 4.7). BPC, BPC(Cl), and

BPAP exposure only induced expression of cyp19a1b, but not the other four biomarkers.

BADGE and BPS exposure did not activate the expression of any of the five biomarkers at any dose (Figure 4.8). In summary, eight bisphenols induced expression of one or more 101 of the estrogenic biomarkers in a dose-dependent manner while BADGE and BPS did not induce any of the biomarkers.

In addition to these five genes, the expression of five other estrogen responsive genes was analyzed after bisphenol exposures, including brain-expressed bndf, liver-expressed fabp10a, kidney-expressed slc6a19b, muscle-expressed atp2a2a, and ubiquitously expressed fkbp5 genes, in an attempt to find more tissue specific biomarkers. However, the alteration of the expression of these genes at this exposure window (4 dpf-6 dpf) was not robust enough for these genes to be estrogenic biomarkers (data not shown). Further investigations about these genes at other exposure windows are necessary.

4.2.3 Anti-estrogenic effects of bisphenols on GFP-expression in Tg(5xERE:GFP) transgenic fish and on expression of estrogenic biomarkers

Since the bisphenols exhibited partial agonistic activities, co-treatment with the bisphenols and the full agonist EE2 was performed to investigate whether the bisphenols could interfere with the EE2-induced activation in the zebrafish larvae. 3 nM EE2 was used for the co-treatment, because this concentration showed the highest antagonistic effect of BPA when co-exposed with EE2 (data not shown). First, GFP expression of the

Tg(5xERE:GFP) fish embryos was analyzed upon co-exposure of EE2 and each bisphenol. Larvae were exposed at 4 dpf to 3 nM EE2 in combination with the highest non-toxic dose of each bisphenol and GFP intensity was quantified using the plate reader at 6 dpf. BPA, BPC, BPC(Cl), BPB, BPE, BPAP, and BADGE all repressed the EE2- induced GFP expression in the transgenic reporter fish (Figure 4.9). BPF, BPAP, and

102

BPS exposure did not show any significant antagonistic effects on the GFP expression.

Co-treatment of EE2 and the ER antagonist ICI 182,780 was used as a positive control

for the anti-estrogenic effect (Figure 4.9).

To further analyze the anti-estrogenic effects of the bisphenols, gene expression levels of

vtg1 and vtg3 were measured using RT-qPCR after exposure to 3nM EE2 combined with

each bisphenol. BPAF, BPB, BPC, and BPC(Cl) showed strong antagonistic effects on

the EE2-induced expression of vtg1 and vtg3, while exposure to BPA, BPE, and BPF

showed lower but still significant repression on EE2-induced expression (Figure 4.10).

BPAP and BPS exposure did not repress the EE2-induced expression of vtg1 and vtg3.

Again, ICI 182,780 co-treatment was used as a positive control and it sufficiently blocked

the induction of vtg1 and vtg3 expression levels by EE2 (Figure 4.10). In conclusion,

exposure to BPA, BPE, BPB, BPAF, BPC, BPC(Cl), and BADGE had anti-estrogenic

effects on both GFP-expression in the transgenic fish and expression of vtgs in wild-type

fish. For BPF, we detected a repression of the induction of vtg1 and vtg3 expression, but

not of GFP expression of the transgenic fish. BPS and BPAP did not have any anti-

estrogenic effects on the transgenic fish, nor on the expression of vtg1 and vtg3.

4.3 Discussion

We assessed the estrogenic and anti-estrogenic effects of BPA and nine BPA derivatives with various structural modifications using zebrafish embryos. Overall, BPA, BPAF,

BPE, BPF, BPB, BPC, BPC(Cl), and BPAP exhibited estrogenic effects but BADGE and

BPS did not. BPA, BPAF, BPE, BPF, BPB, BPC, BPC(Cl), and BADGE showed anti-

103

104

Figure 4.6 EE2 induced vtg1, vtg3, cyp19a1b, f13a1a, and esr1 expression in a dose-dependent manner (A) and the effects were mediated by ERs (B). (A) Dose- response analysis of the relative mRNA expression of the vtg1, vtg3, cyp19a1b, f13a1a, and esr1 in 6 dpf DZ larvae upon EE2 treatment. 4 dpf DZ fish larvae were treated with EE2 ranging from 0.01 nM to 0.1 µM or co-exposed to EE2 for 48 hours. Vehicle only (0.1% DMSO) and ICI treatment groups were used as negative controls. Media were renewed at 5 dpf. Asterisk denotes significant difference (**p<0.01, unpaired student’s t-test compared to the DMSO control). (B) ICI blocked EE2 induced mRNA expression of the vtg1, cyp19a1b, f13a1a, and esr1 in 6 dpf DZ larvae. 4 dpf DZ fish larvae were co-treated with 3nM EE2 and 300 nM ICI for 48 hpf. 3 nM EE2 treatment group was positive control. 300 nM ICI and vehicle only (0.1% DMSO) treatment was used as negative controls. 25 larvae were pooled as one biological sample. Results are presented as relative mRNA expression normalized to 18S. Each experiment was carried out with at least three independent batches of larvae and three technical replicates for each PCR reaction (means ± SD). Asterisk denotes significant difference (**p<0.01, unpaired student’s t-test compared to the EE2 control).

105

A B

C D

106

Figure 4.7 Dose-response analysis of vtg1, vtg3, cyp19a1b, f13a1a, and esr1 expression in 6 dpf DZ larvae exposed to BPA, BPAF, BPE, BPF, and BPB. 4 dpf Tg(5xERE:GFP) fish larvae were treated with BPA (A), BPAF (B), BPE (C), BPF (D), and BPB (E) ranging from 10 nM to 10 µM, or 0.1% DMSO for 48 hours. Media were renewed at 5 dpf. 25 larvae were pooled as one biological sample. Results are presented as relative mRNA expression normalized to 18S. Each experiment was carried out with at least three independent batches of larvae and three technical replicates for each PCR reaction (means ± SD). Asterisk denotes significant difference (**p<0.01, unpaired student’s t-test compared to the DMSO control).

107

A B

BPC (Cl)

C D

108

Figure 4.8 Dose-response analysis of vtg1, vtg3, cyp19a1b, f13a1a, and esr1 expression in 6 dpf DZ larvae exposed to BPC, BPC(Cl), BPAP, BADGE, and BPS. 4 dpf Tg(5xERE:GFP) fish larvae were treated with 10 nM to 10 µM of BPC (A), BPC(Cl) (B), BPAP (C), BADGE (D) and BPS (E), or 0.1% DMSO for 48 hours. Media were renewed at 5 dpf. 25 larvae were pooled as one biological sample. Results are presented as relative mRNA expression normalized to 18S. Each experiment was carried out with at least three independent batches of larvae and three technical replicates for each PCR reaction (means ± SD). Asterisk denotes significant difference (**p<0.01, unpaired student’s t-test compared to the DMSO control).

109

Figure 4.9 GFP fluorescence quantification of Tg(5xERE:GFP) fish larvae following bisphenol treatment in the presence of EE2. 4 dpf Tg(5xERE:GFP) fish larvae were co-treated with EE2 (3 nM) as well as BPA (10 µM), BPE (10 µM), BPF (10 µM), BPB (10 µM), BPS (10 µM), BPAF (3 µM), BPC (3 µM), BPC(Cl) (3 µM), BPAP (3 µM), or BADGE (3 µM) for 48 hours. Vehicle only (0.1% DMSO) and co-exposure of 3 nM EE2 with 300 nM ICI 182,780 were used as controls. 3 nM EE2 treatment was used as a positive control. Media were renewed at 5 dpf. Larvae were transferred to 96-well plates for GFP quantification (5 fish/well as one biological replicate). Results are expressed as relative fluorescence intensity above control (means ± SEM). Asterisk denotes significant difference (**p<0.01, *p<0.05, unpaired student’s t-test compared to the E2 control; ##p<0.01 unpaired student’s t- test compared to the DMSO control). n=8 biological replicates. This experiment was repeated 3 times.

110

Figure 4.10 Anti-estrogenic effects of the 10 bisphenols on expression of vtg1 (A) and vtg3 (B). 4 dpf DZ fish larvae were co-treated with 3 nM EE2 and BPA (10 µM), BPE (10 µM), BPF (10 µM), BPB (10 µM), BPS (10 µM), BPAF (3 µM), BPC (3 µM), BPC(Cl) (3 µM), BPAP (3 µM), or BADGE (3 µM) for 48 hours. Media were renewed at 5 dpf. 25 larvae were pooled as one biological sample. 3 nM EE2 treatment group were used as positive control. Co-exposure of 300nM ICI and 3 nM EE2, as well as vehicle only (0.1% DMSO) groups were used as controls. Results are presented as relative mRNA expression normalized to 18S. Each experiment was carried out with at least three independent batches of larvae and three technical replicates for each PCR reaction (means ± SD). Asterisk denotes significant difference (**p<0.01, *p<0.05, unpaired student’s t-test compared to the EE2 control). 111

estrogenic effects but BPAP and BPS did not. Only BPS did not show either estrogenic or anti-estrogenic effects in our zebrafish assays. BPA, BPAF, BPE, BPF, and BPB showed stronger estrogenic effects than the other bisphenols. Exposure to BPA, BPAF, BPE, BPF, and BPB at the highest non-toxic concentrations activated GFP expression in the transgenic fish in various tissues including liver, heart, brain, eyes, ears, pancreas, and kidney; the highest GFP activation ranged from 50-80% when compared to 3 nM EE2 activation (Table 4.2 and Figure 4.7). In addition, these 5 bisphenols all induced expression of cyp19a1b, whose maximum expression levels ranged from 40-80% when

compared to the maximum expression levels induced by EE2. vtg1, vtg3, and esr1 were also induced by these 5 bisphenols; however their expression levels were relatively low compared to the expression level induced by EE2. f13a1a has never been reported as an estrogenic marker before, but it was highly induced by the 5 bisphenols as well as EE2, indicating its potential as a new biomarker for assessing estrogenic effects (Figure 4.6 and Figure 4.7). BPA has been extensively reported as an ER agonist in both in vitro and in vivo systems (reviewed in [66]), which is in agreement with our data. BPAF has been reported as a full agonist for human ERα but it only displayed extremely weak estrogenic

activity for ERβ in HeLa reporter cells [230]. However, Delfosse and colleagues reported that BPAF activates both ERα and ERβ mediated estrogenic effects when using luciferase

assays in HELN cells [233]. In addition, BPAF exhibits estrogenic effects in MCF-7 cells using ERE-luciferase reporter assay [229] and MCF-7 cell proliferation assay [238]. BPE exhibits estrogenic effects in yeast two-hybrid assays [239]. BPB shows estrogenic effects in yeast two-hybrid assays and MCF-7 reporter cells [229,239]. BPF shows

112

estrogenic potential in yeast two-hybrid assays [239], MCF-7 proliferation assay [238],

and a luciferase reporter assay in MCF7 cells [229]. These results from the in vitro assays

are in agreement with our data on the estrogenicity of BPAF, BPE, BPF, and BPB in

zebrafish embryos.

BPC, BPC(Cl), and BPAP showed very low estrogenic activities in both GFP reporter fish and the biomarker assay (Figures 4.4, 4.5, and 4.8). Exposure to these bisphenols

failed to induce a strong GFP expression in the liver of the GFP reporter fish, which is a

major estrogen responsive organ as shown when the fish are exposed to other bisphenols

(Figure 4.2). Expression of liver-expressed markers vtg1 and vtg3 was not elevated by these bisphenols, which is consistent with the lack of GFP expression in the liver of reporter fish (Figure 4.8). However, exposure to BPC activated GFP expression in the heart, ear, and brain at the highest non-toxic concentration. Exposure to BPC(Cl) activated GFP expression in the brain and ear regions, but not in the heart valves (Figure

4.4). BPAP also activated GFP expression in the brain and ear regions, however, it

induced a much lower activation in the brain than BPC and BPC(Cl) (Figure 4.4). This

expression pattern is distinct from the expression pattern activated by BPA, BPAF, BPE,

BPF, and BPB, since the heart valves seemed to be the most sensitive tissue when the

transgenic fish were exposed to these bisphenols. Consistent with the GFP expression pattern in the reporter fish, only the expression of the brain-expressed marker cyp19a1b

was induced by the exposure of these compounds (Figure 4.8). The estrogenic activities

of BPC(Cl) have been reported, using a cell proliferation assay and luciferase assay in

MCF-7 cells with a stably expressed ERE-driving luciferase expression (MELN cells),

113 which is in line with our data [233]. Moreover, BPC(Cl) shows higher estrogenic activities in the luciferase reporter cell lines bearing esr2a or esr2b than the esr1- expressed cell line (Pinto et al., manuscript in preparation), which may explain its different expression pattern from other bisphenols.

BPS and BADGE did not show any estrogenic activities in either the GFP reporter fish assay or biomarker assay (Figures 4.5 and 4.8). However, BPS has previously been reported to be estrogenic by using a luciferase assay in MELN cells [240], a GFP- expression system in MCF-7 cells [241] and MCF-7 proliferation assay in the E-screen test [242,243]. However, it has also been shown that BPS has no estrogenic activity using the yeast two-hybrid system [242,243] and ERE-luciferase reporter assay in MCF-7 cells

[229]. Grignard and colleagues used in silico metabolism simulation and hypothesized a lower estrogenic activity of BPS metabolites in vivo [240], which might explain the lack of estrogenic effect of BPS in zebrafish. BADGE did not show any estrogenic effect in our studies, which is in agreement with its lack of binding affinity to hERs [232,238,244].

However, it should be noted that its hydrolysis products and derivatives such as dihydrolysed BADGE (BADGE-4OH), chlorohydroxy BADGE (BADGE-2Cl), and bisphenol A diglycidylether dimethacrylate (Bis-GMA) have been shown to be estrogenic using the yeast two-hybrid assay as well as proliferation assays in T47D and MCF-7 breast cancer cells [232,244,245].

Anti-estrogenic effects of these BPA derivatives were also assessed since some of them have been reported to have anti-estrogenic effects in different systems. For example,

BADGE exhibits anti-estrogenic activities in a yeast two-hybrid assay [245]. Weak anti- 114 estrogenic effects were observed with BPB and BPF but not BPAF and BPS in the MCF-

7 luciferase reporter assay [229]. BPAF shows strong anti-estrogenic activity against hERβ but not hERα in the presence of E2 in HeLa reporter cells [230]. BPAF and

BPC(Cl) exhibit anti-estrogenic effects in the presence of E2 in HELN reporter cells

[233]. In our studies, significant antagonistic effects were shown for BPA, BPAF, BPE,

BPB, BPC, BPC(Cl), and BADGE in the transgenic reporter zebrafish and vtgs expression assay (Figures 4.9 and 4.10). In addition, BPF showed anti-estrogenic effects by repressing EE2 activation of vtgs in wild-type zebrafish, but not the GFP expression in the reporter fish (Figure 4.10). Anti-estrogenic effects of the bisphenols in different assays seem to vary more than the estrogenic effects. Molecular mechanisms of antagonism of these compounds need to be further studied.

It has been suggested that the structures of the xenoestrogens may affect their estrogenicity. For example, a critical structure for xenoestrogens to bind to ERs is an unhindered hydroxyl group on an aryl ring and a hydrophobic group attached para to the hydroxyl group [229,246,247]. The structures of the BPA derivatives share a similar backbone of the phenolic hydroxyl group, which meets the key structural requirement for the estrogenic activities. The modifications of the nine BPA derivatives except for BPC and BADGE are mainly at the methylene bridge, whose hydrophobicity has been suggested as an important factor for the estrogenic activities [229]. It was hypothesized that the hydrophobic substituents in the methylene bridge increased the estrogenic activity, suggesting that BPAF and BPB would have stronger estrogenic activities than

BPA. In our studies, BPAF showed relatively higher estrogenic potential than BPA, but

115

BPB did not. Moreover, longer alkyl substituents on the central bridging carbon cause an increased estrogenic potency [238,248]. This is not the case in our studies, though, especially in case of BPB. Furthermore, BPAP showed very weak estrogenic effects, which might be caused by the phenol ring substitution at the methylene bridge, whose steric hindrance reduces the binding affinity of BPAP to the ER ligand-binding pocket.

On the other hand, BPS has a hydrophilic modification between the two phenolic rings, which destroyed the methylene bridge, and the estrogenic effects decreased in our studies, suggesting an important role for the connection between two phenol rings for the estrogenicity. Lastly, BPC contains 3- and 11-methyl groups on the phenyl rings. The substitution at these positions causes enhanced estrogenic and thyroid hormonal activities

[229], but in our studies, BPC exhibited much lower estrogenic activities than BPA in the zebrafish. The structures of the bisphenols may only provide some insight for the ligand binding affinity, however, since factors such as transformation and accumulation of the compounds in vivo need to be taken into considerations for assessing the estrogenicity of the xenoestrogens.

In conclusion, zebrafish embryo is a useful model for assessing the estrogenicity of new compounds. However, there are some limitations to our study. First, we did not detect an effect at environmentally relevant doses. Even our lowest doses are higher than environmentally relevant exposures and they did not show significant effects. Low dose effects have been reported for some environmental chemicals like BPA (reviewed in

[64]). Further studies are necessary to examine low-dose effects of these bisphenol derivatives. Moreover, since some chemicals are toxic to the fish at high concentrations,

116

we chose to use non-toxic concentrations only, but for some compounds like BADGE

and BPS, higher concentrations might be needed to show the estrogenic effect in the in

vivo system due to uptake rate and biotransformation. Also, uptake and metabolism of the

bisphenols need to be tested in future studies. Furthermore, we only exposed the fish at 4 dpf to 6 dpf, however, other exposure windows might exhibit different effects. Lastly, although the GFP-reporter fish showed great potential to assess the estrogenic effects of environmental chemicals, the sensitivity of the fish might need to be further improved since the detected effects all occurred at higher concentrations than the environmentally relevant doses. Knock-down experiments of zfERs can be performed to confirm the tissue specificity of different ER subtypes and rule out crosstalk effects of other signaling pathways like Aryl hydrocarbon receptor (AHR) signaling.

4.4 Conclusions

In summary, we assessed both estrogenic and anti-estrogenic activities of 10 selected bisphenols. BPA, BPAF, BPE, BPB, BPC, and BPC(Cl) all exhibited both estrogenic and anti-estrogenic effects. BPF showed estrogenic effects in both reporter fish and biomarker assay, but only anti-estrogenic effects in the biomarker assay. BPAP only showed weak estrogenic effects, but no anti-estrogenic effects. BADGE showed no estrogenic effects, but strong anti-estrogenic effects. BPS, however, did not show any estrogenic or anti- estrogenic effects in our system, suggesting it might be a better substitute for BPA than the other BPA derivatives. The zebrafish embryo model takes into account the uptake, biotransformation, and elimination of chemicals. In our systems, transgenic reporter fish provided a useful tool to monitor the estrogenic effects in a tissue context and biomarkers 117

provided a complementary tool to confirm the activation of the estrogen signaling in vivo.

BPC, BPC(Cl), and BPAP were shown to induce different GFP expression patterns to

BPA, BPAF, BPE, BPAF, and BPB. The expression of several biomarkers including vtgs, cyp19a1b, esr1, and f13a1a was examined. These markers were found to have different sensitivity to different compounds and they are complementary for the assessment of estrogenic activities. Moreover, f13a1a would be an effective estrogenic biomarker.

These studies provide insight about the endocrine-disrupting activities of 10 bisphenols

and may contribute to the risk assessment of BPA derivatives.

118

Chapter 5. Application of transgenic reporter zebrafish in

high-throughput screening of environmental estrogens and

estrogenic drugs

5.1 Introduction

Zebrafish has been widely used for drug discovery and environmental chemical risk assessment, but these applications were mainly based on the traditional model which involves studying one chemical at a time. Recently, zebrafish has become a promising model for high-throughput screening (HTS) of small molecules. Many features of zebrafish embryos promote their use for HTS, such as low cost, high fecundity, ex-utero and fast embryonic development, small and transparent embryos, evolutionarily conserved signaling pathways with mammalian models, genetically amenable for forward and reverse screenings, and the use of fluorescence reporter fish for cell-type or pathway- specific visualization (reviewed in [77-80]). Using zebrafish in drug screening takes consideration of the ADMET (absorption, distribution, metabolism, excretion, and

toxicity) in the early development process, which could avoid extensive late-stage animal

testing [90]. These advantages greatly promote the use of zebrafish embryos and larvae

for HTS of small molecule libraries.

As environmental chemicals emerge from industry, risk assessment of these chemicals is

highly needed. For prioritization of chemicals for rodent testing, US Environmental

Protection Agency (EPA) has released the ToxCastTM Phase I and Pase II environmental

119

chemical libraries comprising 309 and 767 unique chemicals, respectively

(http://www.epa.gov/ncct/toxcast/chemicals.html). So far, 467 in vitro, high-throughput

screening assays have been performed for assessing chemicals in ToxCastTM Phase I,

including cell-free and cell-based measurements in multiple human primary cells,

engineered cell lines, as well as rat primary hepatocytes [249]. Assay techniques include

reporter gene, competitive binding, and enzyme inhibition assays. Biological targets or

effects include cytotoxicity, cell growth, genotoxicity, enzymatic activity, receptor

binding, ion channels, transcription factor activity and downstream consequences, gene

induction, and high-content imaging of cells [250]. Existing datasets are maintained in

ToxCastDB (http://actor.epa.gov/actor/faces/ToxCastDB/Home.jsp). Endocrine-

disrupting activity assessment organized by the EPA’s Endocrine Disruptor Screening

Program (EDSP) (http://www.epa.gov/endo/) is one of the most important compositions

of this ToxCast project. The endocrine-disrupting activities assessed in this program are

focused on perturbation of estrogen, androgen, and thyroid pathways using in vitro assays

[250], while in vivo assays are very limited. Developmental toxicity of the ToxCastTM

Phase I chemicals has been evaluated using zebrafish embryos with the assessment of

phenotypic endpoints such as lethality, non-hatching, and morphological malformations

[129]. HTS of these chemicals for xenoestrogens in zebrafish would be complementary to

the in vitro assays for prioritization of chemicals.

Zebrafish has also been used for drug discovery pipelines during the past decade; the major uses include screening compounds, drug toxicity, target identification and validation, physiology-based drug discovery, quantitative structure-activity relationship

120

(QSAR), and structure-activity relationship (SAR) study (reviewed in [251], [89]).

Recently, as transgenic reporter fish lines and improved HTS techniques emerged,

zebrafish has started to be used for HTS drug discovery (reviewed in [91,92]). Systems

for loading zebrafish from reservoirs or multi-well plates, as well as positioning and

rotating them for high-speed confocal imaging have been developed, which greatly

facilitates the use of zebrafish for HTS [93]. Automated detection and high-resolution

imaging platforms were also developed for HTS using zebrafish embryos [94-96].

Walker and colleagues developed reporter-based assays involving automated reporter

quantification using several transgenic reporter zebrafish lines [252]. Using a transgenic

zebrafish reporter line generated for analyzing FGF activity, Saydmohammed and

colleagues screened 1,040 compounds and identified two molecules that enhanced FGF signaling in specific areas of the zebrafish brain [253]. Anti-angiogenic compounds were

screened using transgenic zebrafish with fluorescent blood vessels ([97], reviewed in

[254]). Using Tg(dusp6:EGFP) transgenic zebrafish embryos, Molina and colleagues screened chemical libraries including NCI diversity set from NCI/NIH, the Natural

Products library from MicroSource Discovery Systems Incorporation Company, and

Phosphatase targeted set from ChemDiv Incorporation. From these chemicals, they identified a new compound targeting the dual-specificity phosphatase 6 (Dusp6), a component of the fibroblast growth factor (FGF) signaling [98]. These studies indicate that zebrafish model is very useful for HTS in drug development.

Estrogen signaling is well conserved between zebrafish and humans, as discussed in the

previous chapters. Several transgenic reporter zebrafish lines have been generated for

121

estrogen-signaling studies, such as Tg(ERE:Luciferase) fish [255], Tg(ERE-ZVTG1-

EGFP) fish [123], Tg(cyp19a1b-GFP) fish [124], Tg(3xERE:GFP) fish using

Gal4ff/UAS-GFP system [126], and Tg(5xERE:GFP) fish [125]. These reporter fish lines

have great potential for HTS of small molecules, but so far the studies have been limited

to small-scale chemical screening using GFP or luciferase quantification and imaging.

Chen and colleagues tested 9 estrogenic chemicals using adult ERE-ZVTG1-EGFP transgenic fish [123]. Gorelick and colleagues screened 14 chemicals and identified 8

with estrogenic effects using Tg(5xERE:GFP) fish larvae [125], Brion and colleagues

screened 42 compounds and identified 25 with estrogenic effects using cyp19a1b-GFP

transgenic fish embryos [124]. However, a majority of the chemicals in these studies

were redundant and the purposes of these studies were mainly to characterize the

transgenic fish lines. More chemicals need to be screened using these reporter fish to

identify xenoestrogens.

In this chapter, I developed an HTS method to screen for environmental estrogens as well

as clinical drugs with estrogenic activities using embryos of Tg(5xERE:GFP) estrogen

reporter fish. We screened 768 chemicals including 309 chemicals from ToxCastTM Phase

I environmental chemical library, 446 chemicals from NIH clinical collection (NIHCC)

and 14 add-in (cherry-picked) chemicals. Hits from ToxCastTM Phase I library were

further compared to estrogen-related assay results in ToxCastDB from EPA. The hits

from the NIH clinical collection library are classified as estrogenic drugs here.

122

5.2 Results

5.2.1 Optimization of conditions for HTS of chemical libraries using transgenic reporter fish

To use Tg(5xERE:GFP) transgenic reporter fish as an HTS model to screen for xenoestrogens, we first optimized the screening protocol. E2 dose curves were performed to assess the sensitivity of the reporter fish at different exposure windows. Co-exposure of ICI 182,780 (ICI) with E2 was used for antagonizing the E2 activation at these exposure windows. Because the GFP expression was turned on during 24-48 hours (1-2 dpf) after E2 treatment, results had to be analyzed at 2 dpf or later. Five exposure windows were tested, including 1-2 dpf, 1-3 dpf, 1-4 dpf, 3-4 dpf, and 3-5 dpf. Reporter fish embryos were exposed to E2 at 1 dpf or 3 dpf ranging from 0.1 nM to 1 µM, and arrayed in 96-well microplates with 1 embryo/well in 100 µl embryo medium, using 12 replicates for each group. Embryos were anesthetized using eugenol or tricane and GFP expression was observed under the fluorescence microscope and quantified using the plate reader at the end of the exposure. Fish embryos anesthetized by 0.002% eugenol gave more consistent readings and caused less toxic effects than fish anesthetized with tricane (data not shown). Microplates with different colors were used to compare the sensitivity from green, red, blue, and black U-shape wells, and a black plate with clear bottom wells. For GFP expression observation and imaging by microscopy, black microplates with a clear bottom gave the best results with low background noise, however they could not be used for GFP quantification since the clear bottom increased the noise too much in the plate reader. Black U-shape or V-shape bottom microplates yielded the most 123

robust GFP signal in the plate reader compared to plates with other colors; however the

black plates could not be used for imaging (Figure 5.1).

As shown in Figure 5.1, GFP quantification during different exposure windows all

showed monotonic dose-response curves. The GFP intensity was higher at the later

developmental stages, but the variations in signal tended to be larger as the fish grew

older at the higher doses. ICI could abolish the GFP expression induced by E2 during all

treatment windows (data not shown). At 2 dpf, the chorion of embryos and the small size

of the developing liver might explain the low sensitivity (discussed in chapter 3). The

fluorescence background was generally higher in the presence of the chorions in the wells

whether the fish were hatched or not. However, as the fish grew older, the variation

among biological replicates tended to be larger. An interesting finding was that when the

fish embryos were cultured in the black plates, their pigmentation increased compared to

the ones cultured in the clear plates. To avoid this side effect, PTU was added.

In our platform, we combined both GFP quantification using the plate reader and imaging

using fluorescence microscopy. Technically, the plate reader is much faster to screen the compounds than microscopy. It can easily detect strong estrogenic responses. When the fish were dead, the auto-fluorescence doubled the E2-induced signal, while the xenoestrogen-induced signal was normally lower than the E2-induced signal. However, the plate reader protocol had some limitations. First, when the GFP was only expressed in the heart valves and ears but not the liver, the readings were too close to the background to be detected. Secondly, when the fish showed toxic phenotypes such as bent body and heart edema, the auto-fluorescence background went up, sometimes close to the values of 124

the hits. Thus, there was a risk for false positives, depending on the toxicity of the

compounds. In these situations, visual observation and imaging-based analysis using the

black plate with clear bottoms was necessary. Visual observation could detect more hits

especially when the GFP expression was faint and confined. Teratology phenotypes such

as heart edema, bent body, or small head could be recorded for assessing the toxicity of

the compounds. However, manual microscopy is low through-put and time consuming.

Therefore, both the plate reader and microscopy methods would be used as

complimentary methods for screening estrogenic compounds using Tg(5xERE:GFP)

transgenic reporter fish. We chose the exposure window of 3-5 dpf because the

background was low and the induction level was high at this window, and the treatment

was more efficient at a shorter exposure window than at a longer one.

5.2.2 HTS of ToxCastTM Phase I chemical library for environmental estrogens

5.2.2.1 Screening for environmental estrogens using Tg(5xERE:GFP) transgenic

reporter fish

To assess the estrogenic effects of the 309 pesticides in ToxCastTM phase I chemical

library, we exposed Tg(5xERE:GFP) transgenic fish embryos to the chemicals from 3 dpf

until 5 dpf and examined the GFP fluorescence daily. Two fish embryos were distributed

in 100 µl E3 media in 96-well black micro-plates with clear bottom at 3 dpf. 1 µl of 3

mM of the chemical library solutions were added into each well using a multichannel

pipette, which yielded a 30 µM final concentration. Two biological replicates were

examined side by side. E2 and DMSO were used as positive and negative controls,

125

respectively. After 24 h exposure, we found 9 estrogenic hits and 22 compounds with

lethality at 30 µM. However, most of the embryos showed teratological phenotypes such

as bent body, heart edema, or small head at this concentration. After 48 h exposure, there

were 7 more hits, all of which showed GFP expression in the heart valves. 35 more

compounds showed lethal effects to the fish after 48 h exposure. After 72 h exposure, 9

more hits were found showing only GFP expression in the heart valves, but these fish embryos suffered severe heart edema. A majority of the fish were dead after 72 h exposure at 30 µM treatment (Table 5.1 and data not shown). The positive GFP expression, the lethality, and toxicity of the compounds showed high consistency among the 4 biological replicates.

In total, 26 hits were identified with three distinct GFP expression patterns at various exposure windows to the chemical using the transgenic reporter fish. Four chemicals

(HPTE, BPA, MTX and fluazinam) induced a strong GFP expression in various organs such as liver, heart valves, brain, ear region, and eyes, similar with the expression pattern following E2 treatment. Fenhexamid-induced GFP expression in the brain, heart valves, and ear region but not the liver. The other 21 hits only induced GFP expression in the heart valves but not the other organs, indicative of their weak estrogenic effects (Table

5.1).

For the compounds that were lethal for the fish at 30 µM at 4 dpf, lower concentrations were used and were retested using the reporter fish until non-toxic doses were found.

However, no hits were found among these compounds, which might be explained by the

126

A B C

D E

Figure 5.1 E2 dose-response curves at different exposure windows. Panels A-C, Tg(5xERE:GFP) fish embryos were treated from 1 dpf continuously till 2 dpf (A), 3 dpf (B) and 4 dpf (C), and fluorescence was quantified at the end of the treatment. Panels D and E, Tg(5xERE:GFP) fish larvae were treated from 3 dpf continuously till 4 dpf (D) and 5 dpf ( E), and fluorescence was quantified at the end of the treatment. n=12 per treatment group, mean ± SEM. (**indicates p<0.05, student t- test).

127

Table 5.1 Hits from ToxCastTM Phase I chemical library

Exposure Chemical Name CAS No. Time1 Expression Patterns (hours) Strong estrogenic effects liver, heart valves, brain, ear region, eyes, and Bisphenol A 80-05-7 24, 48, 72 pancreas liver, heart valves, brain, ear region, eyes, and HPTE 2971-36-0 24, 48, 72 pancreas Methoxychlor 72-43-5 24 liver, heart valves, and brain Fluazinam 79622-59-6 24 liver, heart valves, and brain Weak estrogenic effects Fenhexamid 126833-17-8 24,48 Heart valves, brain, and ear region Acifluorfen 50594-66-6 72 heart valves Bendiocarb 22781-23-3 72 heart valves Butylate 2008-41-5 72 heart valves Chlorpropham 101-21-3 48 heart valves Clorophene 120-32-1 24, 48 heart valves Cyfluthrin 68359-37-5 48 heart valves Fenarimol 60168-88-9 24, 48 heart valves Fludioxonil 131341-86-1 24, 48 heart valves Flumetralin 62924-70-3 72 heart valves Imazapyr 81334-34-1 72 heart valves Isoxaben 82558-50-7 72 heart valves Maleic hydrazide 123-33-1 48, 72 heart valves Monomethyl 4376-18-5 72 heart valves phthalate Novaluron 116714-46-6 72 heart valves Pendimethalin 40487-42-1 48 heart valves Pirimiphos-methyl 29232-93-7 48 heart valves Propazine 139-40-2 72 heart valves Thiazopyr 117718-60-2 24, 48, 72 heart valves Thiophanate- 23564-05-8 48, 72 heart valves methyl Triadimefon 43121-43-3 48 heart valves Trifluralin 1582-09-8 48 heart valves Note:

1 Exposure windows: 48 hrs: 3 dpf-4 dpf; 72hrs: 3 dpf-5 dpf; 96hrs: 3 dpf-6 dpf;

2 Abbreviation: HPTE: 2,2-Bis(4-hydroxyphenyl)-1,1,1-trichloroethane.

128

fact that the compounds caused lethality at doses that were high enough to show

estrogenic effects in the reporter fish.

5.2.2.2 Comparison of the hits among different estrogen-related assays in

ToxCastDB

We then compared the hits from our in vivo reporter fish assay with hits from the in vitro

assays in ToxCast database (ToxCastDB)

(http://actor.epa.gov/actor/faces/ToxCastDB/Home.jsp). ToxCastDB is the database EPA

maintains to access all data from the ToxCast program which screened 1,000 chemicals in over 500 rapid tests. The 7 assays we compared our hits to also investigated the estrogenic effects of the 309 compounds from ToxCastTM Phase I_v1 using different

platforms. These assays include: Attagene Factorial trans ERα assay

(ATG_ERα_TRANS) and Attagene Factorial cis-ERE assay (ATG_ERE_CIS) HepG2

reporter cell lines; NCGC ERα Agonist Assay (NCGC_ERα_Agonist) and NCGC ERα

Antagonist assay (NCGC_ERα_Antagonist) using HEK293H reporter cell lines;

Novascreen Human ER (NVS_NR_hER), Novascreen Bovine ER (NVS_NR_bER) and

Novascreen Mouse ERα (NVS_NR_mERα) radioactive-binding assays. In total, 90 hits

were found in the ATG_ERα_TRANS assay, 39 hits in the ATG_ERE_CIS assay, 11 hits

in the NCGC_ERα_Agonist, 19 hits in the NCGC_ERα_Antagonist assay, 3 hits in the

NVS_NR_hER assay (HPTE, BPA and MBP), 2 hits in the NVS_NR_bER assay (HPTE

and BPA) and 4 hits in the NVS_NR_mER assay (HPTE, BPA, Clorophene and

Flumetsulam). We generated Venn diagrams to show the overlaps of the hits among

129 different assays using an online Venn diagram tool

(http://bioinfogp.cnb.csic.es/tools/venny/index.html).

As shown in Figure 5.2, there were 7 overlaps when we compared the hits from our reporter fish assay with hits from NCGC_ERα_Agonist assay, 5 overlaps with

NCGC_ERα_Antagonist assay, 10 overlaps with both ATG_ERα_TRANS and

ATG_ERE_CIS assays and 7 overlaps with ATG_ERα_TRANS, ATG_ERE_CIS and

NCGC_ERα_Agonist assays. Detailed overlaps are shown in Table 5.2. Three well- known xenoestrogens MTX, HPTE (metabolite of MTX), and BPA showed up as hits in all of the assays, which showed the comparability of all the different assays, however, additional hits were unique in different assays. The other 4 overlaps were fenhexamid, fludioxonil, flumetralin, and fenarimol whose estrogenic effects call for attention. When compared to NVS human, bovine, and mouse ERs binding assays, we found that flumetsulam from NVS-mERα binding assay and monobutyl phthalate (MBP) from

NVS-hERα binding assay were not overlapping with our hits, while BPA, MTX, and

HPTE overlapped with our hits.

5.2.3 HTS of NIHCC chemical library for estrogenic drugs using Tg(5xERE:GFP) zebrafish

We used the same strategy to screen for estrogenic drugs from the NIHCC clinical collection library, which contains 446 chemicals that have a history of use in human clinical trials for various diseases (Figure 5.3). In addition to the library, we added 13

130

additional chemicals including SERMs, endogenous estrogens (E1 and E3), and a

phytoestrogen (genistein) for reference purposes (Table 5.3).

In contrast to the ToxCastTM Phase I chemicals, the majority of the clinical drugs (more than 90%) were not lethal to the zebrafish embryos within the exposure window (3-5 dpf) at 30 µM, reaffirming their safety since previously they were tested on vertebrate animal models in pre-clinical stages. We found 62 hits out of 460 compounds with five major

GFP expression patterns in the reporter fish. The GFP expression patterns were classified into four groups: (1) all (various organs including liver, heart valves, brain, eyes, ear region), (2) heart valves only, (3) brain and ear region, and (4) lateral line expression. In general, these hits showed much stronger estrogenic effects than the hits from ToxCast library; many of the hits were as strong as endogenous E2. Moreover, most hits activated

GFP expression in the reporter fish after 24 h exposure, which was faster than most of the hits from ToxCastTM Phase I library. We separated the hits from NIHCC library into three

groups based on the classes, use and sources of the compounds: (1) SERMs, steroids and

hormone system-related drugs (Table 5.4), (2) natural products (Table 5.5), and (3) non-

hormone-related drugs (Table 5.6). Table 5.3 shows the add-in compounds which were

also classified into the three main groups.

Hits in group I included known endogenous animal steroids, synthetic steroids (including

SERMs), and drugs that are used for endocrine system related disease (Table 5.3 and 5.4).

For example, this group includes synthetic steroids such as EE2, stanozolol and 19-

nortestosterone, SERMs such as toremifene citrate, propylpyrazole triol (PPT), and

diarylpropionitrile (DPN), as well as the aromatase-inhibiting drug anastrozole used to 131

A B

zf-ERE NCGC-ERα-ag zf-ERE NCGC-ERα-ant

C D

zf-ERE ATG-ERα ATG-ERα ATG-ERE zf-ERE NCGC-ERα-ag

ATG-ERE

Figure 5.2 Venn diagrams showing overlapping environmental estrogens between reporter zebrafish assays and in vitro assays in ToxCastDB. (A) overlapping environmental estrogens between reporter zebrafish assay and NCGC ERα Agonist Assay (NCGC-ERα-ag); (B) overlapping environmental estrogens between reporter zebrafish assay and NCGC ERα Antagonist assay (NCGC-ERα- ant); (C) overlapping environmental estrogens among reporter zebrafish assay, Attagene Factorial tras ERα assay (ATG-ERα) and Attagene Factorial cis ERE assay (ATG-ERE); (D) Overlapping environmental estrogens among reporter zebrafish assay, ATG-ERα assay, ATG-ERE assay and NCGC-ERα-ag Assay.

132

Figure 5.3 NIH Clinical Collection chemical library components (permission authorized, http://www.nihclinicalcollection.com/NCCAnnotatedChartVP.JPG)

133

Table 5.2 Common hits between reporter zebrafish assays and in vitro assays in ToxCastDB

Groups No. of Common elements overlaps zf ERE vs ATG factorial assays zf ERE vs ATG ERα 12 HPTE, Bisphenol A, Methoxychlor, Fenhexamid, Fludioxonil, Flumetralin, Fenarimol, Pirimiphos-methyl, Pendimethalin, Triadimefon, Trifluralin, and Thiazopyr zf ERE vs ATG ERE 11 HPTE, Bisphenol A, Methoxychlor, Fenhexamid, Fludioxonil, Flumetralin, Fenarimol, Pirimiphos-methyl, Pendimethalin, Triadimefon, and Bendiocarb

ATG ERα vs ATG 32 HPTE, Bisphenol A, Methoxychlor, Fenhexamid, Fludioxonil, ERE Flumetralin, Fenarimol, Pirimiphos-methyl, Pendimethalin, Triadimefon , 2-Phenylphenol, Carbaryl, Chlorethoxyfos, Chlorpyrifos oxon, Chlorpyrifos-methyl, Cinmethylin, Diazinon, DBP (Dibutyl phthalate), Dithiopyr, Fenamiphos, Fenthion, Isazofos, Lactofen, Lindane, Malathion, Napropamide, Parathion, PFOA (Perfluorooctanoic acid), Piperonyl butoxide, Pyridaben, Quintozene, and Tribufos zf ERE vs ATG ERα 10 HPTE, Bisphenol A, Methoxychlor, Fenhexamid, Fludioxonil, vs ATT ERE Flumetralin, Fenarimol, Pirimiphos-methyl, Pendimethalin, and Triadimefon zf ERE vs NCGC assays zf ERE vs NCGC 7 HPTE, Bisphenol A, Methoxychlor, Fenhexamid, Fludioxonil, ERα agonist Flumetralin, and Fenarimol zf ERE vs NCGC 5 ERα antagonist HPTE, Bisphenol A, Methoxychlor, Fenarimol, and Clorophene ATG ERα vs NCGC 10 HPTE, Bisphenol A, Methoxychlor, Fenhexamid, Fludioxonil, ESR1 agonist Flumetralin, Fenarimol, Carbaryl, Endosulfan, and Thiobencarb ATG ERE vs NCGC 8 HPTE, Bisphenol A, Methoxychlor, Fenhexamid, Fludioxonil, ERα agonist Flumetralin, Fenarimol, and Carbaryl zf ERE vs ATG factorial assays and NCGC ERα agonist assay zf ERE vs ATG ERα 7 HPTE, Bisphenol A, Methoxychlor, Fenhexamid, Fludioxonil, vs ATG ERE vs Flumetralin, and Fenarimol NCGC ERα agonist

134

Table 5.3 Cherry picked compounds and GFP expression patterns in the reporter fish

Class Doses Chemical Name CAS # (human) Expression patterns Estrone (E1) 53-16-7 Endogenous 1 μM all*

Estriol (E3) 50-27-1 Endogenous 1 μM all*

ERB041 524684-52-4 hERβ 1 μM liver, heart, and selective pancreas

WAY200070 440122-66-7 hERβ 1 μM liver, heart, pancreas selective

FERB033 1111084-78-6 hERβ 1 μM liver and heart selective

Propylpyrazole 263717-53-9 hERα 10 μM ear regions triol (PPT) selective

Diarylpropionitrile 1428-67-7 hERβ 10 μM, Liver (weak), heart, (DPN) selective 1 μM brain, ear regions and pancreas 16alpha-lactone- hERα 10 μM, 1 μM, Liver, heart, brain, ear estradiol selective 100 nM, 10 regions, olfactory area, (16αLE2) nM and pancreas Medicarpin 32383-76-9 hERβ 1 μM, 100 Heart valves selective nM, 10 nM

4OH-Tamoxifene 68047-06-3 hERα 10 μM, 1 uM Toxic but strong GFP selective expression in the deformed liver Genistein 446-72-0 Phytoestrogen 1 μM liver, heart, brain, ear regions, olfactory placode, and pancreas

Tributyltin (TBT) 688-73-3 1 μM Nothing 8-beta-vinyl- hERβ 1 μM Nothing estradiol (8bVE2) selective

ICI 182,780 129453-61-8 Antagonist 1 μM Nothing Note: * All means GFP expression was observed in liver, heart, brain, eyes, ear regions and pancreas.

135 treat breast cancer in postmenopausal women. SERMs showed tissue-specific effects in the reporter fish. For example, exposure to PPT induced GFP expression in the ear region and pancreas of the reporter fish, but only very faint GFP expression in the liver.

Exposure to FERB033 induced GFP expression in the liver and heart valves and very strong expression in the olfactory bulbs of the reporter fish. Exposure to ERB041 induced

GFP expression in the liver, heart valves, pancreas, and some spotted regions in the head which are different from the typical expression regions. Exposure to WAY200070 induced GFP expression in the liver, heart valves, and pancreas (Figure 5.4).

Group II included hits from NIHCC as well as the added-in compounds genistein and medicarpin (Table 5.5). These compounds are natural products originating from bacteria, fungi, and plants (but not animals). For example, dactinomycine can be isolated from soil bacteria and is used as an antibiotic and chemotherapeutic drug. Zeranol, derived from fungi, is a known non-steroidal estrogen agonist. Maltol, brucine, icariin, and secoisolariciresinol can also be found in plants. Exposure to zeranol, brucine, maltol, and secoisolariciresinol induced similar GFP expression patterns as E2-induced GFP expression in the reporter fish. Taxifolin, a flavanol, only induced GFP expression in the ear regions. Resveratrol and dactinomycin only induced GFP expression in the heart valves (Table 5.5). When the reporter fish were exposed icariin, a primary active component of epimedium extracts, GFP expression was shown in the liver and heart, however, line-shaped fluorescence was also observed in the head and gut of the embryos.

Similar fluorescence was also detected by exposure to three flavonoids (hyperoside, rutin, or isoquercitrin) (Figure 5.5). However, exposure of these four chemicals to wild-type

136

DZ fish embryos showed similar fluorescence, indicating auto-fluorescence of the

chemicals themselves (data not shown).

Group III included non-hormone related synthetic drugs, such as anti-inflammatory drugs, antidepressant drugs, antidopaminergic drugs, antibiotics, and anticonvulsant drugs

(Table 5.6). These compounds have not been used as clinical estrogens, however, many of them showed very strong estrogenic effects, inducing GFP expression in the reporter fish in a similar pattern as E2 does (for example, the anti-inflammatory balsalazide, antidepressant citalopram hydrobromide, antibiotic enrofloxacin, and antimetabolite

raltitrexed). Some compounds showed weaker estrogenic effects, inducing GFP

expression in the heart valves but not liver in the reporter fish, such as the anti-diabetic

drug rosiglitazone maleate, antiplatelet drug ticlopidine hydrochloride (TICLID), and

anti-inflammatory drug naproxen sodium. Strikingly, we found the compound RU 24969 activated a unique GFP-expression pattern in the reporter fish that has never been

observed before. In addition to liver, brain, and heart valves, the GFP was expressed

throughout the lateral line including both the anterior lateral line in the head and posterior lateral line on the trunk and tail (Figure 5.5). In summary, using Tg(5xERE:GFP), I

efficiently found 62 hits from the 446 chemicals contained in the NIH clinical collection drug library and 13 added-in chemicals. The GFP-expression patterns activated by these estrogenic drugs are different. Further studies will be performed for potential SERMs.

137

5.3 Discussion

5.3.1 Hits from ToxCastTM Phase I chemical library (environmental estrogens)

We used Tg(5xERE:GFP) fish to screen for xenoestrogens in ToxCastTM Phase I pesticide library. Only BPA, MTX, HPTE, and fluazinam showed strong estrogenic effects, inducing GFP expression in the liver, brain, and heart valves of the reporter fish.

BPA has been extensively reported as a xenoestrogen (reviewed in [66]). MTX has been reported to exhibit estrogenic activity in a mouse uterotrophic assay, but not in mouse ER binding and HeLa ER cell transcriptional activation assays [256]. Gaido and colleagues showed that HPTE, the metabolite of MTX exhibits estrogenic activity in vitro [74], which explained why MTX is estrogenic in vivo but not in vitro. When incubated with human hepatic microsomes, the estrogenic potency of MTX was highly increased in the yeast cell assay, indicating that the metabolites including HPTE are more potent than

MTX itself [257]. Beresford and colleagues confirmed the higher potency of HPTE than

MTX [258]. Moreover, Gaido and colleagues reported that HPTE was preferably an ERα agonist with antagonist activity on ERβ in human hepatoma cells [74,259]. The remaining hits from ToxCastTM Phase I showed weak estrogenic effects, inducing GFP expression only in the heart valves at 30 µM. Some of them only induced GFP expression after continuous treatment for several days. The different expression patterns might result from different ER selectivity of these compounds.

We compared our hits from the transgenic fish assay with the estrogen related in vitro assays in ToxCastDB, however, only 7 overlapping chemicals were found across all

138

Table 5.4 Estrogenic drugs I: endogenous estrogen, steroids and steroidgenesis related drugs

Pubchem_ID Chemical Name Expression patterns Estrogens Liver, heart, brain, eyes, ear regions, and 46386902 E2 pancreas Liver, heart, brain, eyes, ear regions, and 46386858 EE2 pancreas 46386573 Ethylestrenol Liver, heart, brain, and eyes 46386940 Mestranol Liver, heart, brain, and eyes 46386677 Levonorgestrel Liver, heart, brain, and eyes 1,3,5(10)-estratrien-3- 46386974 ol-17-one sulphate, Liver, heart, brain, and eyes sodium salt Androgens 46501387 Testosterone Liver, heart, brain, and eyes 46387028 Dehydroepiandrosterone Liver, heart, brain, and eyes 46386759 Mestanolone Liver, heart, brain, and eyes 46386580 Methylandrostenediol Liver, heart, brain, eyes, and ear regions 46386739 Methyltestosterone Liver, heart, brain, and eyes 46386608 Stanozolol Liver, heart, brain, and eyes 46386825 Tibolone Liver, heart, brain, and eyes 46386764 19-nortestosterone Liver, heart, brain, and eyes

Others

46386818 Goserelin acetate Liver, heart, brain, and eyes

46386543 Anastrozole Liver, heart, brain, and eyes

46386786 Toremifene citrate Liver, heart, brain, and eyes

139

Table 5.5 Estrogenic drugs II: Natural products

Pubchem_ID Chemical Names Expression patterns 46386627 Zeranol Liver, heart, brain, and eyes 46386962 Resveratrol Heart 46386883 Dactinomycin Heart 46386939 Brucine Liver, heart, brain, and eyes 46386578 Icariin Liver, heart, brain, eyes and ear regions 46386945 Maltol Liver, heart, brain, and eyes 46386735 Secoisolariciresinol Liver, heart, brain, and eyes 46386773 Taxifolin-(+/-) Ear regions

140

Table 5.6 Estrogenic drugs III: Synthetic drugs

Pubchem_ID Chemical Names Expression patterns 46386887 6-aminoindazole Liver, heart, brain, and eyes 46386828 Balsalazide Liver, heart, brain, and eyes 46386942 Benactyzine hydrochloride Liver, heart, brain, and eyes 46386617 Citalopram hydrobromide Liver, heart, brain, and eyes 46386821 Doxapram hydrochloride Liver, heart, brain, and eyes 46386863 Droperidol Liver, heart, brain, and eyes 46386889 Enrofloxacin Liver, heart, brain, and eyes 46386904 Fluperlapine Liver, heart, brain, and eyes Liver, heart, brain, eyes, and ear 46386710 Gabexate mesylate regions 46386806 Haloperidol Liver, heart, brain, and eyes 46386566 Irbesartan Liver, heart, brain, and eyes 46386834 Ketotifen fumarate Liver, heart, brain, and eyes 46386860 Metronidazole Liver, heart, brain, and eyes 46386949 Milnacipran hydrochloride Liver, heart, brain, and eyes 46386964 Nimodipine Liver, heart, brain, and eyes 46386961 Oxyphenonium bromide Liver, heart, brain, and eyes 46386841 Paroxetine Liver, heart, brain, and eyes 46386823 Raltitrexed Liver, heart, brain, and eyes 46387003 Rimcazole Liver, heart, brain, and eyes 46386966 Rolitetracycline Liver, heart, brain, and eyes 46386628 Rosiglitazone maleate Heart and ear regions 46386912 Stiripentol Liver, heart, brain, and eyes 46386594 Tropisetronâ hydrochloride Liver, heart, brain, and eyes 46386948 Naproxen sodium Heart 46386576 Ticlopidine Hydrochloride Heart Liver, heart, ear regions, and lateral 46387004 RU 24969 line system

141

Figure 5.4 GFP expression patterns of 5dpf old Tg(5xERE:GFP)zebrafish larvae exposed to 1 µM SERMs. Indicated selectivity of the ligands is based on in vitro zebrafish assays for zebrafish ER selectivity (Pinto et al., manuscript in preparation). PPT: propylpyrazole triol; FEB: FERB033; ERB: ERB041; WAY: WAY200070. e: ear region; of: olfactory bulbs.

142

Figure 5.5 GFP fluorescence of Tg(5xERE: GFP) zebrafish larvae exposed to EE2, isoquercitrin and RU24969. 10 nM EE2 induced GFP expression in the liver, heart, brain, eyes, ear regions, and pancreas in 5 dpf embryos (A). GFP fluorescence observed in the 30 µM isoquercitrin treated fish at 5 dpf due to the auto-fluorescence of isoquercitrin chemical solution (B). 30 µM RU 24969 induced GFP expression in the liver, heart valves, brain, ear regions and lateral line system in 4 dpf embryos (C). Brightness and contrast was adjusted for picture quality.

143

assays, and most of the hits were unique in different assays. Among the seven common

hits across all four estrogenic assays (Figure 5.2), three of them are well-known xenoestrogens MTX, HPTE, and BPA. This confirmed the reliability of the

Tg(5xERE:GFP) reporter fish. Moreover, fenhexamid, fludioxonil, flumetralin, and fenarimol also showed estrogenic activities in both in vitro assays and our reporter fish

assay. Fenhexamid and fludioxonil are commonly used antifungal agents for fruits and

vegetables [260]. They have been reported to show anti-androgenic and anti-estrogenic

activity in engineered human breast cancer cells [260,261]. Furthermore, fenhexamid

inhibits an enzyme, C3-ketoreductase, which is involved in ergosterol biosynthesis in

fungi [262]. However, fenhexamid, fludioxonil, and flumetralin did not cause substantial

reproductive toxicity in rats [263]. Flumetralin and fenarimol were shown to be

estrogenic in vivo for the first time in our reporter fish assay. Modes of action of these

compounds need to be further studied.

Some pesticides like endosulfan, parathion, and carbaryl are hits in vitro, but in our

reporter fish system, they did not exhibit any estrogenic effect. Endosulfan did not show

uterotrophic or other endocrine-related endpoints in immature female rats [264].

However, endosulfan has been reported to induce vtg1 expression in zebrafish embryos

[265]. This might due to different exposure windows, or different sensitivity of the two

assays. Parathion and carbaryl do not have any estrogenic activity in MCF-7 and MDA-

MB-231 cells [266]. However, anti-estrogenic activities of carbaryl and endosulfan have

been shown in HELN cell lines [267]. These conflicting results need to be further

clarified.

144

The inconsistencies of different assays might result from species, tissue, or cell-type differences across different assays. Most of the in vitro assays used engineered human cell lines, which might be substantially different from our ERE reporter zebrafish system.

Chemical uptake and metabolism are important factors too. Some chemicals might be metabolically activated, while others might be metabolically deactivated in vivo. In addition, toxicity of the compounds in the zebrafish might cause some false negatives, since some chemicals showed estrogenicity at concentrations in vitro that are lethal in vivo. Moreover, some compounds might be androgenic and could be converted to estrogenic compounds in the presence of AroB (encoded by cyp19a1b) in zebrafish. The compounds might also act on the AhR, and affect ER function through direct receptor binding or multiple indirect receptor pathways via AhR [173,268]. Further studies in the same system would be necessary to confirm their estrogenicity.

5.3.2 Hits from NIHCC (Clinical estrogens and phytoestrogens)

We used Tg(5xERE:GFP) reporter fish to screen for clinical estrogens from the NIHCC library. Hits were separated into three groups: endogenous and clinical steroids as well as hormone system-related drugs, natural products, and non-hormone-related drugs.

In group I, besides the estrogens, androgens also activated GFP expression in the reporter fish (Table 5.3 and Table 5.4). This is probably because androgen in the presence of

AroB can be converted into estrogen in vivo, which subsequently binds to ERs and activates GFP expression in the reporter fish. Testosterone and the non-aromatizable

DHT have been reported to induce GFP expression in the Tg(cyp19a1b-GFP) zebrafish

145

embryos [124]. The DHT effect also involves its conversion to the metabolite 5α- androstane-3β, 17β-diol (3β-Adiol) with known estrogenic activity [269]. Exposure to

most of the SERMs in group I, such as toremifene citrate, PPT, and DPN, strongly

activated GFP expression in various tissues (Table 5.3). The hERβ-specific agonist 8-

beta-vinyl-estradiol (8β-VE2) did not show estrogenic effect at 1 μM in the reporter fish,

however, it has been shown to exhibit zfER-mediated transcriptional activation in

engineered HELN cell line (data not shown). This might be because the concentration

tested in the fish was not high enough to activate GFP in the transgenic fish.

Group II compounds were the natural products. Genistein is a well-known phytoestrogen.

It has been shown to activate GFP expression in the Tg(5xERE:GFP) reporter fish [125]

and Tg(cyp19a1b-GFP) reporter fish [124]. Of our hits, zeranol, resveratrol, icariin, and

taxifolin are all phytoestrogens. Zeranol, genistein, and taxifolin have been shown to

exhibit estrogenic effects in mouse uterotrophic assay [270]. Resveratrol, produced by

some spermatophytes, has also been reported to be estrogenic [271]. Icariin exposure

does not induce estrogenic effects in itself [272,273], but its metabolites, icaritin and

desmethylicaritin, significantly increase proliferation in the MCF-7 cell assay [273,274].

These results are in line with our reporter fish assay.

Drugs in Class III were originally used for non-hormone-related purposes and most of

them have not been reported as xenoestrogens before. The fact that there were several

hits in this group suggests that caution should be taken regarding the side-effects of these

drugs (Table 5.6). On the other hand, these drugs might be new ER ligands, which can be

used for hormone-related disease. Since these drugs are already approved by FDA, new 146

therapeutic use of them can save time and money in the drug discovery pipeline. RU

24969, a hit from NIHCC in our reporter fish assay might be a potential drug for a second

therapeutic usage. It is an agonist for both 5-hydroxytryptamine receptor 1A (5-

HT1A) and 5-hydroxytryptamine receptor 1B (5-HT1B), but preferentially an agonist for

5-HT1B. However, it has been reported that estrogen interacts with 5-HT receptors

(reviewed in [275,276]). Moreover, RU24969 can stimulate sexual behavior in female

rats via increasing 5-HT activity [277]. In our reporter fish assay, exposure to RU 24969

induced GFP expression in the lateral line system (Figure 5.5), which has been used to

screen drugs for hearing loss [278]. Furthermore, lateral line expression has been shown

to be correlated with Esr2a in zebrafish larvae [109]. Therefore, this drug might be an

ERβ selective ligand. Further studies are needed to illustrate the relationship between RU

24969 and ER signaling.

5.3.3 Tg(5xERE:GFP) fish as a HTS model for screening for environmental and

estrogenic drugs

Tg(5xERE:GFP) reporter fish have been shown to be a useful tool for HTS of

environmental and estrogenic drugs. Not only developmental toxicity, but also estrogenic

activities of the chemicals can be assessed using this reporter fish. Tissue-specific and

time-dependent effects of xenoestrogens could be analyzed from the primary screening.

Compared to the traditional screening model of engineered cell lines expressing separate

ER subtypes, the reporter fish model is more efficient as it can be used to screen for full

ER ligands and selective ER ligands all at once. In addition, different expression patterns might reflect ER subtype selectivity. Emerging high-throughput imaging techniques will 147

also facilitate chemical screening for large small molecule libraries using this reporter

fish.

However this model has its limitations. First, this is only an ERE-based assay. Some

other nuclear receptors might bind to EREs and cause non-specific effects, and vice versa,

ERE-independent signaling, such as ER signaling through AP1 sites will not be detected

with this fish line. Other chemicals might activate phosphorylation of ERs, resulting in

activation of cytoplasmic signal transduction pathways that eventually results in an ERE-

independent genomic response. These pathways will also be missed using the ERE

reporter fish. In vitro binding assays to each ER subtype and functional assays of the hits

need to be further studied. All in all, reporter fish serve as a complementary model for screening environmental estrogens and estrogenic drugs.

148

Chapter 6. Conclusions and Future Perspectives

This dissertation maps out estrogen signaling pathways during zebrafish embryonic

development and suggests candidate biomarkers for detection of xenoestrogenic exposure

to zebrafish embryos. The estrogenic and anti-estrogenic effects of a class of bisphenols

were assessed using these biomarkers as well as a transgenic reporter fish,

Tg(5xERE:GFP). Using the reporter fish, clinical drugs and environmental pollutants

from small molecular chemical libraries were screened for estrogenic activity in a high- throughput manner, and several new compounds with estrogenic activity were identified.

To better understand estrogen signaling in zebrafish, we performed transcriptome

analysis at four developmental stages. The main findings of this analysis were that the expression of the identified estrogen responsive genes to a large extent was distinct at 1, 2,

3 and 4 dpf. Despite this fact, the GO biological functional groups were relatively similar among the four time points. This could be explained by the fact that distinct signaling pathways are involved in the development of different functional organs at different times, while the overall estrogenic effect promoting development is conserved. Similar to mammals, the estrogen signaling was distinctly regulated in different tissues and organs in zebrafish. Brain, liver, heart, and pancreas were major estrogen responsive organs in developing zebrafish embryos. Ovary and testis categories were enriched from the differentially expressed genes in the embryos using online tissue enrichment analysis tool

at ZFIN, but the gonad differentiation in zebrafish is not complete until 40 dpf [155].

This might be because the gonadal progenitor genes are turned on in the early

149 developmental stages. GFP expression in these organs was observed in the adult

Tg(5xERE:GFP) transgenic fish which is in line with the enrichment analysis. A majority of the estrogen target tissues was confirmed both in embryonic and adult stages using the transgenic reporter fish Tg(5xERE:GFP). When compared with E2-treated adult fish datasets published on GEO, most differentially expressed genes from the embryonic stages were distinct from the ones in the adults. Only the expression of a few genes was co-regulated, indicating their potential to be used as biomarkers in both embryos and adult fish. One such common gene was f13a1a, the expression of which was highly induced at all developmental stages and adult. f13a1a encodes factor VIII, a key factor in the blood coagulation pathway. We further found that the expression of several other factors in this pathway was regulated by estrogen.

The findings from the zebrafish models can be extrapolated to mammalian systems. The differentially expressed genes from our microarray studies showed a high overlap with the mammalian estrogen target genes as discussed in chapter 3. In addition, using the mammalian homologues of the zebrafish differentially expressed genes, we found that the enriched GO biological functional groups are conserved with human estrogen regulated functional groups (chapter 3, [133]). Moreover, the estrogenic responsive tissues during zebrafish development also shared similarities with those of rodents and humans, including liver, brain and pancreas. However, there are also differences between zebrafish and mammalian estrogen signaling. Due to the evolutionary genome duplication, one human gene homologue often matches several zebrafish gene subtypes, but these subtypes in zebrafish might have evolved to play different roles mediating

150 estrogen signaling. For example, in humans there is one gene for cyp19a1, but in fish there are two: cyp19a1a, which is expressed in the ovary, and cyp19a1b, which is expressed in the brain. These genes play different roles in response to estrogens. It was reported that exposure to EE2, 4-Nonylphenol (NP) and BPA all down-regulated the expression of cyp19a1a in rare minnow juveniles. However, for the expression of cyp19a1b, it was increased by EE2, decreased by BPA but not affected by 4-NP. [236].

There might also be tissue differences between mammals and fish; a major estrogenic responsive tissue in fish is the heart valve as detected by our reporter zebrafish. The heart valve has not been reported to be regulated by estrogens in the mammalian systems.

Zebrafish larva was shown to be an effective tool for studying estrogenic disruption.

Estrogenic and anti-estrogenic effects of 10 bisphenols were investigated using both wild type and transgenic zebrafish larvae. Exposure to eight bisphenols resulted in tissue specific and dose dependent estrogenic effects in zebrafish larvae. Exposure to one compound, BADGE, induced anti-estrogenic effects but not estrogenic effects in zebrafish larvae. The only compound that was inert in the fish was BPS; neither estrogenic nor anti-estrogenic effects were detected after exposure to it. Thus, I conclude that most of the selected bisphenols in our studies that could serve as substitutes for BPA have the capacity to interfere with normal estrogenic signaling in zebrafish, and thus might not be environmentally safe. Moreover, we analyzed whether the expression of f13a1a was induced by the bisphenols and found that it was efficiently induced by the bisphenols that had estrogenic activity, but not by the ones that lacked estrogenic activity.

This finding confirms that f13a1a works as a new biomarker for estrogenicity. Further

151

studies will be done to analyze tissue distribution of f13a1a expression and its sensitivity

in detecting exposure to different xenoestrogens. Molecular mechanisms of the

estrogenicity of the bisphenols also need to be further analyzed, as they might have

different modes of action at the molecular level.

Tg(5×ERE:GFP) transgenic reporter zebrafish were used as a high-throughput model to

screen for environmental estrogens and estrogenic drugs. The urge for risk assessment of

the emerging environmental chemicals requires efficient and low-cost models for HTS.

Zebrafish embryos as a small teleost model fit into this context perfectly. Based on our

genomic studies, estrogen signaling is highly conserved between zebrafish and mammals,

but the fish model is much more efficient for HTS. Furthermore, as an in vivo model,

zebrafish screening takes into consideration the physiological context of drug metabolism,

and thus, it serves as a complementary model to in vitro models for risk assessment of

environmental chemicals. An additional advantage of the reporter fish is that tissue-

specific effects can be detected in the primary screening.

Using Tg(5×ERE:GFP) transgenic reporter zebrafish, we screened for environmental

estrogens from the ToxCastTM Phase I pesticide library. Comparison of hits for

estrogenicity among the pesticides in ToxCastTM Phase I library from the reporter fish to the in vitro models in ToxRefDB showed some overlap. The difference between the hits may result from the distinct metabolism of chemicals between the cell lines and the zebrafish larvae, the lack of metabolism in certain cell lines, or the distribution of compounds in different tissues in zebrafish larvae. Some chemicals might be activated, whereas others might be de-activated in the zebrafish liver. More in vivo models are 152

needed to fully assess the estrogenicity of the pesticides to compare the species

differences and further confirm the metabolism of chemicals. Another use of the HTS

based on transgenic reporter fish is to screen for estrogenic activity of clinical drugs.

Using the NIH clinical drug library to start with, we aimed at finding new use for these

drugs, for example as SERMs. In this thesis, the primary screening of the estrogenic drugs is included. Estrogenic drugs from the NIH clinical collection showed tissue specific GFP expression patterns in the transgenic fish. ER-subtype selectivity may be one cause of the different GFP expression patterns, since the domination of ER subtypes might vary in different estrogen responsive tissues. Those compounds that activated unique GFP expression patterns might be potential SERMs. Further studies of the human

ER subtype selectivity need to be performed and structure-activity relationships (SAR) of

the chemicals can be analyzed.

In summary, the zebrafish embryo is a useful model for studying estrogen signaling and

estrogenic disruption. GFP reporter fish show great potential as a high-throughput in vivo

model for screening environmental estrogens and estrogenic drugs. More small-molecule

libraries can be screened for both toxicological and pharmaceutical purposes using this

model. However, to further screen for SERMs using this model, tissue-specific

expression of the three zebrafish ER subtypes needs to be more extensively characterized.

Furthermore, ER-subtype knock-out ERE reporter fish can be generated to study the

function of different ERs in zebrafish and to screen for subtype specific compounds.

153

References

1. Nelson LR, Bulun SE (2001) Estrogen production and action. J Am Acad Dermatol 45: S116-124. 2. Boon WC, Chow JD, Simpson ER (2010) The multiple roles of estrogens and the enzyme aromatase. Prog Brain Res 181: 209-232. 3. Nilsson S, Gustafsson JA (2011) Estrogen receptors: therapies targeted to receptor subtypes. Clin Pharmacol Ther 89: 44-55. 4. Prossnitz ER, Barton M (2011) The G-protein-coupled estrogen receptor GPER in health and disease. Nat Rev Endocrinol 7: 715-726. 5. Gronemeyer H, Gustafsson JA, Laudet V (2004) Principles for modulation of the nuclear receptor superfamily. Nat Rev Drug Discov 3: 950-964. 6. Heldring N, Pike A, Andersson S, Matthews J, Cheng G, et al. (2007) Estrogen receptors: how do they signal and what are their targets. Physiol Rev 87: 905-931. 7. Nilsson S, Makela S, Treuter E, Tujague M, Thomsen J, et al. (2001) Mechanisms of estrogen action. Physiol Rev 81: 1535-1565. 8. Drummond AE, Fuller PJ (2010) The importance of ERbeta signalling in the ovary. J Endocrinol 205: 15-23. 9. Nilsson S, Koehler KF, Gustafsson JA (2011) Development of subtype-selective oestrogen receptor-based therapeutics. Nat Rev Drug Discov 10: 778-792. 10. Taylor AH, Al-Azzawi F (2000) Immunolocalisation of oestrogen receptor beta in human tissues. J Mol Endocrinol 24: 145-155. 11. Feng Y, Gregor P (1997) Cloning of a novel member of the G protein-coupled receptor family related to peptide receptors. Biochem Biophys Res Commun 231: 651-654. 12. Carmeci C, Thompson DA, Ring HZ, Francke U, Weigel RJ (1997) Identification of a gene (GPR30) with homology to the G-protein-coupled receptor superfamily associated with estrogen receptor expression in breast cancer. Genomics 45: 607- 617. 13. Kvingedal AM, Smeland EB (1997) A novel putative G-protein-coupled receptor expressed in lung, heart and lymphoid tissue. FEBS Lett 407: 59-62. 14. Owman C, Blay P, Nilsson C, Lolait SJ (1996) Cloning of human cDNA encoding a novel heptahelix receptor expressed in Burkitt's lymphoma and widely distributed in brain and peripheral tissues. Biochem Biophys Res Commun 228: 285-292. 15. O'Dowd BF, Nguyen T, Marchese A, Cheng R, Lynch KR, et al. (1998) Discovery of three novel G-protein-coupled receptor genes. Genomics 47: 310-313. 16. Revankar CM, Cimino DF, Sklar LA, Arterburn JB, Prossnitz ER (2005) A transmembrane intracellular estrogen receptor mediates rapid cell signaling. Science 307: 1625-1630. 17. Filardo EJ, Quinn JA, Bland KI, Frackelton AR, Jr. (2000) Estrogen-induced activation of Erk-1 and Erk-2 requires the G protein-coupled receptor homolog, GPR30, and occurs via trans-activation of the epidermal growth factor receptor through release of HB-EGF. Mol Endocrinol 14: 1649-1660.

154

18. Olde B, Leeb-Lundberg LM (2009) GPR30/GPER1: searching for a role in estrogen physiology. Trends Endocrinol Metab 20: 409-416. 19. Kanda N, Watanabe S (2003) 17beta-estradiol inhibits oxidative stress-induced apoptosis in keratinocytes by promoting Bcl-2 expression. J Invest Dermatol 121: 1500-1509. 20. Kanda N, Watanabe S (2003) 17Beta-estradiol enhances the production of nerve growth factor in THP-1-derived macrophages or peripheral blood monocyte- derived macrophages. J Invest Dermatol 121: 771-780. 21. Kanda N, Watanabe S (2004) 17beta-estradiol stimulates the growth of human keratinocytes by inducing cyclin D2 expression. J Invest Dermatol 123: 319-328. 22. le Maire A, Bourguet W, Balaguer P (2010) A structural view of nuclear hormone receptor: endocrine disruptor interactions. Cell Mol Life Sci 67: 1219-1237. 23. Grigsby JG, Parvathaneni K, Almanza MA, Botello AM, Mondragon AA, et al. (2011) Effects of tamoxifen versus raloxifene on retinal capillary endothelial cell proliferation. J Ocul Pharmacol Ther 27: 225-233. 24. Jordan VC (2001) Selective estrogen receptor modulation: a personal perspective. Cancer Res 61: 5683-5687. 25. Meltser I, Tahera Y, Simpson E, Hultcrantz M, Charitidi K, et al. (2008) Estrogen receptor beta protects against acoustic trauma in mice. J Clin Invest 118: 1563- 1570. 26. Harris HA, Katzenellenbogen JA, Katzenellenbogen BS (2002) Characterization of the biological roles of the estrogen receptors, ERalpha and ERbeta, in estrogen target tissues in vivo through the use of an ERalpha-selective ligand. Endocrinology 143: 4172-4177. 27. Brzozowski AM, Pike AC, Dauter Z, Hubbard RE, Bonn T, et al. (1997) Molecular basis of agonism and antagonism in the oestrogen receptor. Nature 389: 753-758. 28. Nolte RT, Wisely GB, Westin S, Cobb JE, Lambert MH, et al. (1998) Ligand binding and co-activator assembly of the peroxisome proliferator-activated receptor- gamma. Nature 395: 137-143. 29. Shiau AK, Barstad D, Loria PM, Cheng L, Kushner PJ, et al. (1998) The structural basis of estrogen receptor/coactivator recognition and the antagonism of this interaction by tamoxifen. Cell 95: 927-937. 30. Webb P, Nguyen P, Kushner PJ (2003) Differential SERM effects on corepressor binding dictate ERalpha activity in vivo. J Biol Chem 278: 6912-6920. 31. Bjornstrom L, Sjoberg M (2005) Mechanisms of estrogen receptor signaling: convergence of genomic and nongenomic actions on target genes. Mol Endocrinol 19: 833-842. 32. Powell E, Xu W (2008) Intermolecular interactions identify ligand-selective activity of estrogen receptor alpha/beta dimers. Proc Natl Acad Sci U S A 105: 19012- 19017. 33. Monroe DG, Secreto FJ, Subramaniam M, Getz BJ, Khosla S, et al. (2005) Estrogen receptor alpha and beta heterodimers exert unique effects on estrogen- and tamoxifen-dependent gene expression in human U2OS osteosarcoma cells. Mol Endocrinol 19: 1555-1568.

155

34. Charn TH, Liu ET, Chang EC, Lee YK, Katzenellenbogen JA, et al. (2010) Genome- wide dynamics of chromatin binding of estrogen receptors alpha and beta: mutual restriction and competitive site selection. Mol Endocrinol 24: 47-59. 35. Safe S, Kim K (2008) Non-classical genomic estrogen receptor (ER)/specificity protein and ER/activating protein-1 signaling pathways. J Mol Endocrinol 41: 263-275. 36. Hall JM, McDonnell DP (2005) Coregulators in nuclear estrogen receptor action: from concept to therapeutic targeting. Mol Interv 5: 343-357. 37. Hammes SR, Levin ER (2007) Extranuclear steroid receptors: nature and actions. Endocr Rev 28: 726-741. 38. Chambliss KL, Wu Q, Oltmann S, Konaniah ES, Umetani M, et al. (2010) Non- nuclear estrogen receptor alpha signaling promotes cardiovascular protection but not uterine or breast cancer growth in mice. J Clin Invest 120: 2319-2330. 39. Lu Q, Pallas DC, Surks HK, Baur WE, Mendelsohn ME, et al. (2004) Striatin assembles a membrane signaling complex necessary for rapid, nongenomic activation of endothelial NO synthase by estrogen receptor alpha. Proc Natl Acad Sci U S A 101: 17126-17131. 40. Watters JJ, Campbell JS, Cunningham MJ, Krebs EG, Dorsa DM (1997) Rapid membrane effects of steroids in neuroblastoma cells: effects of estrogen on mitogen activated protein kinase signalling cascade and c-fos immediate early gene transcription. Endocrinology 138: 4030-4033. 41. Endoh H, Sasaki H, Maruyama K, Takeyama K, Waga I, et al. (1997) Rapid activation of MAP kinase by estrogen in the bone cell line. Biochem Biophys Res Commun 235: 99-102. 42. Chen Z, Yuhanna IS, Galcheva-Gargova Z, Karas RH, Mendelsohn ME, et al. (1999) Estrogen receptor alpha mediates the nongenomic activation of endothelial nitric oxide synthase by estrogen. J Clin Invest 103: 401-406. 43. Migliaccio A, Di Domenico M, Castoria G, de Falco A, Bontempo P, et al. (1996) Tyrosine kinase/p21ras/MAP-kinase pathway activation by estradiol-receptor complex in MCF-7 cells. EMBO J 15: 1292-1300. 44. Madak-Erdogan Z, Kieser KJ, Kim SH, Komm B, Katzenellenbogen JA, et al. (2008) Nuclear and extranuclear pathway inputs in the regulation of global gene expression by estrogen receptors. Mol Endocrinol 22: 2116-2127. 45. Marino M, Acconcia F, Bresciani F, Weisz A, Trentalance A (2002) Distinct nongenomic signal transduction pathways controlled by 17beta-estradiol regulate DNA synthesis and cyclin D(1) gene transcription in HepG2 cells. Mol Biol Cell 13: 3720-3729. 46. Castoria G, Migliaccio A, Bilancio A, Di Domenico M, de Falco A, et al. (2001) PI3- kinase in concert with Src promotes the S-phase entry of oestradiol-stimulated MCF-7 cells. EMBO J 20: 6050-6059. 47. Razandi M, Pedram A, Greene GL, Levin ER (1999) Cell membrane and nuclear estrogen receptors (ERs) originate from a single transcript: studies of ERalpha and ERbeta expressed in Chinese hamster ovary cells. Mol Endocrinol 13: 307- 319.

156

48. Pappas TC, Gametchu B, Watson CS (1995) Membrane estrogen receptors identified by multiple antibody labeling and impeded-ligand binding. FASEB J 9: 404-410. 49. Toran-Allerand CD, Guan X, MacLusky NJ, Horvath TL, Diano S, et al. (2002) ER- X: a novel, plasma membrane-associated, putative estrogen receptor that is regulated during development and after ischemic brain injury. J Neurosci 22: 8391-8401. 50. Acconcia F, Ascenzi P, Fabozzi G, Visca P, Marino M (2004) S-palmitoylation modulates human estrogen receptor-alpha functions. Biochem Biophys Res Commun 316: 878-883. 51. Filardo EJ, Quinn JA, Frackelton AR, Jr., Bland KI (2002) Estrogen action via the G protein-coupled receptor, GPR30: stimulation of adenylyl cyclase and cAMP- mediated attenuation of the epidermal growth factor receptor-to-MAPK signaling axis. Mol Endocrinol 16: 70-84. 52. Zhang L, Blackman BE, Schonemann MD, Zogovic-Kapsalis T, Pan X, et al. (2010) Estrogen receptor beta-selective agonists stimulate calcium oscillations in human and mouse embryonic stem cell-derived neurons. PLoS One 5: e11791. 53. Harrington WR, Kim SH, Funk CC, Madak-Erdogan Z, Schiff R, et al. (2006) Estrogen dendrimer conjugates that preferentially activate extranuclear, nongenomic versus genomic pathways of estrogen action. Mol Endocrinol 20: 491-502. 54. Duan R, Porter W, Safe S (1998) Estrogen-induced c-fos protooncogene expression in MCF-7 human breast cancer cells: role of estrogen receptor Sp1 complex formation. Endocrinology 139: 1981-1990. 55. Duan R, Xie W, Burghardt RC, Safe S (2001) Estrogen receptor-mediated activation of the serum response element in MCF-7 cells through MAPK-dependent phosphorylation of Elk-1. J Biol Chem 276: 11590-11598. 56. Duan R, Xie W, Li X, McDougal A, Safe S (2002) Estrogen regulation of c-fos gene expression through phosphatidylinositol-3-kinase-dependent activation of serum response factor in MCF-7 breast cancer cells. Biochem Biophys Res Commun 294: 384-394. 57. Castro-Rivera E, Samudio I, Safe S (2001) Estrogen regulation of cyclin D1 gene expression in ZR-75 breast cancer cells involves multiple enhancer elements. J Biol Chem 276: 30853-30861. 58. Li C, Briggs MR, Ahlborn TE, Kraemer FB, Liu J (2001) Requirement of Sp1 and estrogen receptor alpha interaction in 17beta-estradiol-mediated transcriptional activation of the low density lipoprotein receptor gene expression. Endocrinology 142: 1546-1553. 59. Distefano E, Marino M, Gillette JA, Hanstein B, Pallottini V, et al. (2002) Role of tyrosine kinase signaling in estrogen-induced LDL receptor gene expression in HepG2 cells. Biochim Biophys Acta 1580: 145-149. 60. Colborn T, vom Saal FS, Soto AM (1993) Developmental effects of endocrine- disrupting chemicals in wildlife and humans. Environ Health Perspect 101: 378- 384.

157

61. McLachlan JA (2001) Environmental signaling: what embryos and evolution teach us about endocrine disrupting chemicals. Endocr Rev 22: 319-341. 62. Casals-Casas C, Desvergne B (2011) Endocrine disruptors: from endocrine to metabolic disruption. Annu Rev Physiol 73: 135-162. 63. Tabb MM, Blumberg B (2006) New modes of action for endocrine-disrupting chemicals. Mol Endocrinol 20: 475-482. 64. Vandenberg LN, Colborn T, Hayes TB, Heindel JJ, Jacobs DR, Jr., et al. (2012) Hormones and endocrine-disrupting chemicals: low-dose effects and nonmonotonic dose responses. Endocr Rev 33: 378-455. 65. Rhomberg LR, Goodman JE (2012) Low-dose effects and nonmonotonic dose- responses of endocrine disrupting chemicals: has the case been made? Regul Toxicol Pharmacol 64: 130-133. 66. Vandenberg LN, Maffini MV, Sonnenschein C, Rubin BS, Soto AM (2009) Bisphenol-A and the great divide: a review of controversies in the field of endocrine disruption. Endocr Rev 30: 75-95. 67. Ryokkynen A, Kayhko UR, Mustonen AM, Kukkonen JV, Nieminen P (2005) Multigenerational exposure to phytosterols in the mouse. Reprod Toxicol 19: 535- 540. 68. Darney S, Fowler B, Grandjean P, Heindel J, Mattison D, et al. (2011) Prenatal Programming and Toxicity II (PPTOX II): role of environmental stressors in the developmental origins of disease. Reprod Toxicol 31: 271. 69. Skinner MK, Manikkam M, Guerrero-Bosagna C (2011) Epigenetic transgenerational actions of endocrine disruptors. Reprod Toxicol 31: 337-343. 70. Shanle EK, Xu W (2011) Endocrine disrupting chemicals targeting estrogen receptor signaling: identification and mechanisms of action. Chem Res Toxicol 24: 6-19. 71. Hannah R, D'Aco VJ, Anderson PD, Buzby ME, Caldwell DJ, et al. (2009) Exposure assessment of 17alpha-ethinylestradiol in surface waters of the United States and Europe. Environ Toxicol Chem 28: 2725-2732. 72. Park KJ, Muller CT, Markman S, Swinscow-Hall O, Pascoe D, et al. (2009) Detection of endocrine disrupting chemicals in aerial invertebrates at sewage treatment works. Chemosphere 77: 1459-1464. 73. Kuiper GG, Lemmen JG, Carlsson B, Corton JC, Safe SH, et al. (1998) Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta. Endocrinology 139: 4252-4263. 74. Gaido KW, Leonard LS, Maness SC, Hall JM, McDonnell DP, et al. (1999) Differential interaction of the methoxychlor metabolite 2,2-bis-(p- hydroxyphenyl)-1,1,1-trichloroethane with estrogen receptors alpha and beta. Endocrinology 140: 5746-5753. 75. Vandenberg LN, Chahoud I, Heindel JJ, Padmanabhan V, Paumgartten FJ, et al. (2010) Urinary, circulating, and tissue biomonitoring studies indicate widespread exposure to bisphenol A. Environ Health Perspect 118: 1055-1070. 76. Lawson ND, Wolfe SA (2011) Forward and reverse genetic approaches for the analysis of vertebrate development in the zebrafish. Dev Cell 21: 48-64.

158

77. Sipes NS, Padilla S, Knudsen TB (2011) Zebrafish: as an integrative model for twenty-first century toxicity testing. Birth Defects Res C Embryo Today 93: 256- 267. 78. Scholz S, Fischer S, Gundel U, Kuster E, Luckenbach T, et al. (2008) The zebrafish embryo model in environmental risk assessment--applications beyond acute toxicity testing. Environ Sci Pollut Res Int 15: 394-404. 79. Augustine-Rauch K, Zhang CX, Panzica-Kelly JM (2010) In vitro developmental toxicology assays: A review of the state of the science of rodent and zebrafish whole embryo culture and embryonic stem cell assays. Birth Defects Res C Embryo Today 90: 87-98. 80. McCollum CW, Ducharme NA, Bondesson M, Gustafsson JA (2011) Developmental toxicity screening in zebrafish. Birth Defects Res C Embryo Today 93: 67-114. 81. Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF (1995) Stages of embryonic development of the zebrafish. Dev Dyn 203: 253-310. 82. Tao T, Peng J (2009) Liver development in zebrafish (Danio rerio). J Genet Genomics 36: 325-334. 83. Farooq M, Sulochana KN, Pan X, To J, Sheng D, et al. (2008) Histone deacetylase 3 (hdac3) is specifically required for liver development in zebrafish. Dev Biol 317: 336-353. 84. Tiso N, Moro E, Argenton F (2009) Zebrafish pancreas development. Mol Cell Endocrinol 312: 24-30. 85. Barros RP, Gustafsson JA (2011) Estrogen receptors and the metabolic network. Cell Metab 14: 289-299. 86. Paskova V, Hilscherova K, Blaha L (2011) Teratogenicity and embryotoxicity in aquatic organisms after pesticide exposure and the role of oxidative stress. Rev Environ Contam Toxicol 211: 25-61. 87. Incardona JP, Collier TK, Scholz NL (2004) Defects in cardiac function precede morphological abnormalities in fish embryos exposed to polycyclic aromatic hydrocarbons. Toxicol Appl Pharmacol 196: 191-205. 88. Chen J, Carney SA, Peterson RE, Heideman W (2008) Comparative genomics identifies genes mediating cardiotoxicity in the embryonic zebrafish heart. Physiol Genomics 33: 148-158. 89. Chakraborty C, Hsu CH, Wen ZH, Lin CS, Agoramoorthy G (2009) Zebrafish: a complete animal model for in vivo drug discovery and development. Curr Drug Metab 10: 116-124. 90. Williams CH, Hong CC (2011) Multi-step usage of in vivo models during rational drug design and discovery. Int J Mol Sci 12: 2262-2274. 91. Lessman CA (2011) The developing zebrafish (Danio rerio): a vertebrate model for high-throughput screening of chemical libraries. Birth Defects Res C Embryo Today 93: 268-280. 92. Zon LI, Peterson RT (2005) In vivo drug discovery in the zebrafish. Nat Rev Drug Discov 4: 35-44. 93. Pardo-Martin C, Chang TY, Koo BK, Gilleland CL, Wasserman SC, et al. (2010) High-throughput in vivo vertebrate screening. Nat Methods 7: 634-636.

159

94. Peravali R, Gehrig J, Giselbrecht S, Lutjohann DS, Hadzhiev Y, et al. (2011) Automated feature detection and imaging for high-resolution screening of zebrafish embryos. Biotechniques 50: 319-324. 95. Vogt A, Cholewinski A, Shen X, Nelson SG, Lazo JS, et al. (2009) Automated image-based phenotypic analysis in zebrafish embryos. Dev Dyn 238: 656-663. 96. Vogt A, Codore H, Day BW, Hukriede NA, Tsang M (2010) Development of automated imaging and analysis for zebrafish chemical screens. J Vis Exp. 97. Tran TC, Sneed B, Haider J, Blavo D, White A, et al. (2007) Automated, quantitative screening assay for antiangiogenic compounds using transgenic zebrafish. Cancer Res 67: 11386-11392. 98. Molina G, Vogt A, Bakan A, Dai W, Queiroz de Oliveira P, et al. (2009) Zebrafish chemical screening reveals an inhibitor of Dusp6 that expands cardiac cell lineages. Nat Chem Biol 5: 680-687. 99. Bardet PL, Horard B, Robinson-Rechavi M, Laudet V, Vanacker JM (2002) Characterization of oestrogen receptors in zebrafish (Danio rerio). J Mol Endocrinol 28: 153-163. 100. Menuet A, Pellegrini E, Anglade I, Blaise O, Laudet V, et al. (2002) Molecular characterization of three estrogen receptor forms in zebrafish: binding characteristics, transactivation properties, and tissue distributions. Biol Reprod 66: 1881-1892. 101. Costache AD, Pullela PK, Kasha P, Tomasiewicz H, Sem DS (2005) Homology- modeled ligand-binding domains of zebrafish estrogen receptors alpha, beta1, and beta2: from in silico to in vivo studies of estrogen interactions in Danio rerio as a model system. Mol Endocrinol 19: 2979-2990. 102. Liu X, Zhu P, Sham KW, Yuen JM, Xie C, et al. (2009) Identification of a membrane estrogen receptor in zebrafish with homology to mammalian GPER and its high expression in early germ cells of the testis. Biol Reprod 80: 1253- 1261. 103. Chandrasekar G, Archer A, Gustafsson JA, Andersson Lendahl M (2010) Levels of 17beta-estradiol receptors expressed in embryonic and adult zebrafish following in vivo treatment of natural or synthetic ligands. PLoS One 5: e9678. 104. Tingaud-Sequeira A, Andre M, Forgue J, Barthe C, Babin PJ (2004) Expression patterns of three estrogen receptor genes during zebrafish (Danio rerio) development: evidence for high expression in neuromasts. Gene Expr Patterns 4: 561-568. 105. Mouriec K, Lareyre JJ, Tong SK, Le Page Y, Vaillant C, et al. (2009) Early regulation of brain aromatase (cyp19a1b) by estrogen receptors during zebrafish development. Dev Dyn 238: 2641-2651. 106. Lassiter CS, Kelley B, Linney E (2002) Genomic structure and embryonic expression of estrogen receptor beta a (ERbetaa) in zebrafish (Danio rerio). Gene 299: 141-151. 107. Celeghin A, Benato F, Pikulkaew S, Rabbane MG, Colombo L, et al. (2011) The knockdown of the maternal estrogen receptor 2a (esr2a) mRNA affects embryo

160

transcript contents and larval development in zebrafish. Gen Comp Endocrinol 172: 120-129. 108. Bertrand S, Thisse B, Tavares R, Sachs L, Chaumot A, et al. (2007) Unexpected novel relational links uncovered by extensive developmental profiling of nuclear receptor expression. PLoS Genet 3: e188. 109. Froehlicher M, Liedtke A, Groh K, Lopez-Schier H, Neuhauss SC, et al. (2009) Estrogen receptor subtype beta2 is involved in neuromast development in zebrafish (Danio rerio) larvae. Dev Biol 330: 32-43. 110. Menuet A, Le Page Y, Torres O, Kern L, Kah O, et al. (2004) Analysis of the estrogen regulation of the zebrafish estrogen receptor (ER) reveals distinct effects of ERalpha, ERbeta1 and ERbeta2. J Mol Endocrinol 32: 975-986. 111. Segner H (2009) Zebrafish (Danio rerio) as a model organism for investigating endocrine disruption. Comp Biochem Physiol C Toxicol Pharmacol 149: 187-195. 112. Meng X, Bartholomew C, Craft JA (2010) Differential expression of vitellogenin and oestrogen receptor genes in the liver of zebrafish, Danio rerio. Anal Bioanal Chem 396: 625-630. 113. Kausch U, Alberti M, Haindl S, Budczies J, Hock B (2008) Biomarkers for exposure to estrogenic compounds: gene expression analysis in zebrafish (Danio rerio). Environ Toxicol 23: 15-24. 114. Muncke J, Eggen RI (2006) Vitellogenin 1 mRNA as an early molecular biomarker for endocrine disruption in developing zebrafish (Danio rerio). Environ Toxicol Chem 25: 2734-2741. 115. Wang J, Shi X, Du Y, Zhou B (2011) Effects of xenoestrogens on the expression of vitellogenin (vtg) and cytochrome P450 aromatase (cyp19a and b) genes in zebrafish (Danio rerio) larvae. J Environ Sci Health A Tox Hazard Subst Environ Eng 46: 960-967. 116. Henry TB, McPherson JT, Rogers ED, Heah TP, Hawkins SA, et al. (2009) Changes in the relative expression pattern of multiple vitellogenin genes in adult male and larval zebrafish exposed to exogenous estrogens. Comp Biochem Physiol A Mol Integr Physiol 154: 119-126. 117. Chiang EF, Yan YL, Guiguen Y, Postlethwait J, Chung B (2001) Two Cyp19 (P450 aromatase) genes on duplicated zebrafish chromosomes are expressed in ovary or brain. Mol Biol Evol 18: 542-550. 118. Kishida M, Callard GV (2001) Distinct cytochrome P450 aromatase isoforms in zebrafish (Danio rerio) brain and ovary are differentially programmed and estrogen regulated during early development. Endocrinology 142: 740-750. 119. Kishida M, McLellan M, Miranda JA, Callard GV (2001) Estrogen and xenoestrogens upregulate the brain aromatase isoform (P450aromB) and perturb markers of early development in zebrafish (Danio rerio). Comp Biochem Physiol B Biochem Mol Biol 129: 261-268. 120. Cheshenko K, Pakdel F, Segner H, Kah O, Eggen RI (2008) Interference of endocrine disrupting chemicals with aromatase CYP19 expression or activity, and consequences for reproduction of teleost fish. Gen Comp Endocrinol 155: 31-62.

161

121. Legler J, Zeinstra LM, Schuitemaker F, Lanser PH, Bogerd J, et al. (2002) Comparison of in vivo and in vitro reporter gene assays for short-term screening of estrogenic activity. Environ Sci Technol 36: 4410-4415. 122. Bogers R, Mutsaerds E, Druke J, De Roode DF, Murk AJ, et al. (2006) Estrogenic endpoints in fish early life-stage tests: luciferase and vitellogenin induction in estrogen-responsive transgenic zebrafish. Environ Toxicol Chem 25: 241-247. 123. Chen H, Hu J, Yang J, Wang Y, Xu H, et al. (2010) Generation of a fluorescent transgenic zebrafish for detection of environmental estrogens. Aquat Toxicol 96: 53-61. 124. Brion F, Le Page Y, Piccini B, Cardoso O, Tong SK, et al. (2012) Screening estrogenic activities of chemicals or mixtures in vivo using transgenic (cyp19a1b- GFP) zebrafish embryos. PLoS One 7: e36069. 125. Gorelick DA, Halpern ME (2011) Visualization of estrogen receptor transcriptional activation in zebrafish. Endocrinology 152: 2690-2703. 126. Lee O, Takesono A, Tada M, Tyler CR, Kudoh T (2012) Biosensor zebrafish provide new insights into potential health effects of environmental estrogens. Environ Health Perspect 120: 990-996. 127. Kallivretaki E, Eggen R, Neuhauss S, Alberti M, Kausch U, et al. (2006) Aromatase in zebrafish: a potential target for endocrine disrupting chemicals. Mar Environ Res 62 Suppl: S187-190. 128. Houck KA, Dix DJ, Judson RS, Kavlock RJ, Yang J, et al. (2009) Profiling bioactivity of the ToxCast chemical library using BioMAP primary human cell systems. J Biomol Screen 14: 1054-1066. 129. Padilla S, Corum D, Padnos B, Hunter DL, Beam A, et al. (2012) Zebrafish developmental screening of the ToxCast Phase I chemical library. Reprod Toxicol 33: 174-187. 130. Fushimi S, Wada N, Nohno T, Tomita M, Saijoh K, et al. (2009) 17beta-Estradiol inhibits chondrogenesis in the skull development of zebrafish embryos. Aquat Toxicol 95: 292-298. 131. Nelson BP, Henriet RP, Holt AW, Bopp KC, Houser AP, et al. (2008) The role of estrogen in the developmental appearance of sensory-motor behaviors in the zebrafish (Danio rerio): the characterization of the "listless" model. Brain Res 1222: 118-128. 132. Hamad A, Kluk M, Fox J, Park M, Turner JE (2007) The effects of aromatase inhibitors and selective estrogen receptor modulators on eye development in the zebrafish (Danio rerio). Curr Eye Res 32: 819-827. 133. Lam SH, Lee SG, Lin CY, Thomsen JS, Fu PY, et al. (2011) Molecular conservation of estrogen-response associated with cell cycle regulation, hormonal carcinogenesis and cancer in zebrafish and human cancer cell lines. BMC Med Genomics 4: 41. 134. Levi L, Pekarski I, Gutman E, Fortina P, Hyslop T, et al. (2009) Revealing genes associated with vitellogenesis in the liver of the zebrafish (Danio rerio) by transcriptome profiling. BMC Genomics 10: 141.

162

135. Ruggeri B, Ubaldi M, Lourdusamy A, Soverchia L, Ciccocioppo R, et al. (2008) Variation of the genetic expression pattern after exposure to estradiol-17beta and 4-nonylphenol in male zebrafish (Danio rerio). Gen Comp Endocrinol 158: 138- 144. 136. Tong Y, Shan T, Poh YK, Yan T, Wang H, et al. (2004) Molecular cloning of zebrafish and medaka vitellogenin genes and comparison of their expression in response to 17beta-estradiol. Gene 328: 25-36. 137. Tong S-K, Mouriec K, Kuo M-W, Pellegrini E, Gueguen M-M, et al. (2009) A cyp19a1b-gfp (aromatase B) transgenic zebrafish line that expresses GFP in radial glial cells. genesis 47: 67-73. 138. Muncke J, Junghans M, Eggen RI (2007) Testing estrogenicity of known and novel (xeno-)estrogens in the MolDarT using developing zebrafish (Danio rerio). Environ Toxicol 22: 185-193. 139. Wang H, Tan JT, Emelyanov A, Korzh V, Gong Z (2005) Hepatic and extrahepatic expression of vitellogenin genes in the zebrafish, Danio rerio. Gene 356: 91-100. 140. Uzumcu M, Kuhn PE, Marano JE, Armenti AE, Passantino L (2006) Early postnatal methoxychlor exposure inhibits folliculogenesis and stimulates anti-Mullerian hormone production in the rat ovary. J Endocrinol 191: 549-558. 141. Armenti AE, Zama AM, Passantino L, Uzumcu M (2008) Developmental methoxychlor exposure affects multiple reproductive parameters and ovarian folliculogenesis and gene expression in adult rats. Toxicol Appl Pharmacol 233: 286-296. 142. Hashimoto S, Miwa M, Akasofu K, Nishida E (1991) Changes in 40 serum proteins of post-menopausal women. Maturitas 13: 23-33. 143. Lee O, Takesono A, Tada M, Tyler CR, Kudoh T (2012) Biosensor zebrafish provide new insights into potential health effects of environmental estrogens. Environ Health Perspect 120: 990-996. 144. Fan X, Xu H, Warner M, Gustafsson JA (2010) ERbeta in CNS: new roles in development and function. Prog Brain Res 181: 233-250. 145. Le Saux M, Estrada-Camarena E, Di Paolo T (2006) Selective estrogen receptor- alpha but not -beta agonist treatment modulates brain alpha-amino-3-hydroxy-5- methyl-4-isoxazolepropionic acid receptors. J Neurosci Res 84: 1076-1084. 146. Liu F, Day M, Muniz LC, Bitran D, Arias R, et al. (2008) Activation of estrogen receptor-beta regulates hippocampal synaptic plasticity and improves memory. Nat Neurosci 11: 334-343. 147. Ram Murthy A, Purushottam M, Kiran Kumar HB, Vallikiran M, Krishna N, et al. (2012) Gender-specific association of TSNAX/DISC1 locus for schizophrenia and bipolar affective disorder in South Indian population. J Hum Genet 57:523-530. 148. Jorgensen A, Morthorst JE, Andersen O, Rasmussen LJ, Bjerregaard P (2008) Expression profiles for six zebrafish genes during gonadal sex differentiation. Reprod Biol Endocrinol 6: 25. 149. Hinfray N, Palluel O, Turies C, Cousin C, Porcher JM, et al. (2006) Brain and gonadal aromatase as potential targets of endocrine disrupting chemicals in a model species, the zebrafish (Danio rerio). Environ Toxicol 21: 332-337.

163

150. Pellegrini E, Menuet A, Lethimonier C, Adrio F, Gueguen MM, et al. (2005) Relationships between aromatase and estrogen receptors in the brain of teleost fish. Gen Comp Endocrinol 142: 60-66. 151. Stoilov I, Rezaie T, Jansson I, Schenkman JB, Sarfarazi M (2004) Expression of cytochrome P4501b1 (Cyp1b1) during early murine development. Mol Vis 10: 629-636. 152. Jansson I, Stoilov I, Sarfarazi M, Schenkman JB (2001) Effect of two mutations of human CYP1B1, G61E and R469W, on stability and endogenous steroid substrate metabolism. Pharmacogenetics 11: 793-801. 153. Jennings BL, Anderson LJ, Estes AM, Yaghini FA, Fang XR, et al. (2012) Cytochrome P450 1B1 contributes to renal dysfunction and damage caused by angiotensin II in mice. Hypertension 59: 348-354. 154. Clelland E, Peng C (2009) Endocrine/paracrine control of zebrafish ovarian development. Mol Cell Endocrinol 312: 42-52. 155. Orban L, Sreenivasan R, Olsson PE (2009) Long and winding roads: testis differentiation in zebrafish. Mol Cell Endocrinol 312: 35-41. 156. Germana A, Laura R, Montalbano G, Guerrera MC, Amato V, et al. (2010) Expression of brain-derived neurotrophic factor and TrkB in the lateral line system of zebrafish during development. Cell Mol Neurobiol 30: 787-793. 157. Jones DR, Schmidt RJ, Pickard RT, Foxworthy PS, Eacho PI (2002) Estrogen receptor-mediated repression of human hepatic lipase gene transcription. J Lipid Res 43: 383-391. 158. Dubey RK, Imthurn B, Barton M, Jackson EK (2005) Vascular consequences of menopause and hormone therapy: importance of timing of treatment and type of estrogen. Cardiovasc Res 66: 295-306. 159. Lamon-Fava S, Ordovas JM, Schaefer EJ (1999) Estrogen increases apolipoprotein (apo) A-I secretion in hep G2 cells by modulating transcription of the apo A-I gene promoter. Arterioscler Thromb Vasc Biol 19: 2960-2965. 160. Lamon-Fava S, Postfai B, Diffenderfer M, DeLuca C, O'Connor J, Jr., et al. (2006) Role of the estrogen and progestin in hormonal replacement therapy on apolipoprotein A-I kinetics in postmenopausal women. Arterioscler Thromb Vasc Biol 26: 385-391. 161. Kushwaha RS, Foster DM, Murthy VN, Carey KD, Bernard MG (1990) Metabolic regulation of apoproteins of high-density lipoproteins by estrogen and progesterone in the baboon (Papio sp). Metabolism 39: 544-552. 162. Dormire SL (2009) The potential role of glucose transport changes in hot flash physiology: a hypothesis. Biol Res Nurs 10: 241-247. 163. Medina RA, Meneses AM, Vera JC, Guzman C, Nualart F, et al. (2004) Differential regulation of glucose transporter expression by estrogen and progesterone in Ishikawa endometrial cancer cells. J Endocrinol 182: 467-478. 164. Hart CD, Flozak AS, Simmons RA (1998) Modulation of glucose transport in fetal rat lung: a sexual dimorphism. Am J Respir Cell Mol Biol 19: 63-70.

164

165. Xu H, Uno JK, Inouye M, Xu L, Drees JB, et al. (2003) Regulation of intestinal NaPi-IIb cotransporter gene expression by estrogen. Am J Physiol Gastrointest Liver Physiol 285: G1317-1324. 166. Chen G, Yang X, Alber S, Shusterman V, Salama G (2011) Regional genomic regulation of cardiac sodium-calcium exchanger by oestrogen. J Physiol 589: 1061-1080. 167. Tazebay UH, Wapnir IL, Levy O, Dohan O, Zuckier LS, et al. (2000) The mammary gland iodide transporter is expressed during lactation and in breast cancer. Nat Med 6: 871-878. 168. Dohan O, De la Vieja A, Carrasco N (2006) Hydrocortisone and purinergic signaling stimulate sodium/iodide symporter (NIS)-mediated iodide transport in breast cancer cells. Mol Endocrinol 20: 1121-1137. 169. Chung JK (2002) Sodium iodide symporter: its role in nuclear medicine. J Nucl Med 43: 1188-1200. 170. De Geeter F, Goethals L (2011) Unilateral breast uptake of 99mTc-pertechnetate. Hell J Nucl Med 14: 76-77. 171. Sanchez-Criado JE, de Las Mulas JM, Bellido C, Navarro VM, Aguilar R, et al. (2006) Gonadotropin-secreting cells in ovariectomized rats treated with different oestrogen receptor ligands: a modulatory role for ERbeta in the gonadotrope? J Endocrinol 188: 167-177. 172. Trukhacheva E, Lin Z, Reierstad S, Cheng YH, Milad M, et al. (2009) Estrogen receptor (ER) beta regulates ERalpha expression in stromal cells derived from ovarian endometriosis. J Clin Endocrinol Metab 94: 615-622. 173. Beischlag TV, Perdew GH (2005) ER alpha-AHR-ARNT protein-protein interactions mediate estradiol-dependent transrepression of dioxin-inducible gene transcription. J Biol Chem 280: 21607-21611. 174. Elgort MG, Zou A, Marschke KB, Allegretto EA (1996) Estrogen and estrogen receptor antagonists stimulate transcription from the human retinoic acid receptor- alpha 1 promoter via a novel sequence. Mol Endocrinol 10: 477-487. 175. Han QX, Allegretto EA, Shao ZM, Kute TE, Ordonez J, et al. (1997) Elevated expression of retinoic acid receptor-alpha (RAR alpha) in estrogen-receptor- positive breast carcinomas as detected by immunohistochemistry. Diagn Mol Pathol 6: 42-48. 176. Deng L, Shipley GL, Loose-Mitchell DS, Stancel GM, Broaddus R, et al. (2003) Coordinate regulation of the production and signaling of retinoic acid by estrogen in the human endometrium. J Clin Endocrinol Metab 88: 2157-2163. 177. Lai K, Harnish DC, Evans MJ (2003) Estrogen receptor alpha regulates expression of the orphan receptor small heterodimer partner. J Biol Chem 278: 36418-36429. 178. Saxena D, Escamilla-Hernandez R, Little-Ihrig L, Zeleznik AJ (2007) Liver receptor homolog-1 and steroidogenic factor-1 have similar actions on rat granulosa cell steroidogenesis. Endocrinology 148: 726-734. 179. Tsai MS, Bogart DF, Li P, Mehmi I, Lupu R (2002) Expression and regulation of Cyr61 in human breast cancer cell lines. Oncogene 21: 964-973.

165

180. Xie D, Miller CW, O'Kelly J, Nakachi K, Sakashita A, et al. (2001) Breast cancer. Cyr61 is overexpressed, estrogen-inducible, and associated with more advanced disease. J Biol Chem 276: 14187-14194. 181. Thongngarm T, Jenkins JK, Ndebele K, McMurray RW (2003) Estrogen and progesterone modulate monocyte cell cycle progression and apoptosis. Am J Reprod Immunol 49: 129-138. 182. Wilkins HR, Doucet K, Duke V, Morra A, Johnson N (2010) Estrogen prevents sustained COLO-205 human colon cancer cell growth by inducing apoptosis, decreasing c-myb protein, and decreasing transcription of the anti-apoptotic protein bcl-2. Tumour Biol 31: 16-22. 183. Pantschenko AG, Zhang W, Nahounou M, McCarthy MB, Stover ML, et al. (2005) Effect of osteoblast-targeted expression of bcl-2 in bone: differential response in male and female mice. J Bone Miner Res 20: 1414-1429. 184. Singer CA, Rogers KL, Dorsa DM (1998) Modulation of Bcl-2 expression: a potential component of estrogen protection in NT2 neurons. Neuroreport 9: 2565- 2568. 185. Kim R, Tanabe K, Emi M, Uchida Y, Osaki A, et al. (2005) Rationale for sequential tamoxifen and anticancer drugs in adjuvant setting for patients with node- and receptor-positive breast cancer. Int J Oncol 26: 1025-1031. 186. Deb S, Amin S, Imir AG, Yilmaz MB, Suzuki T, et al. (2004) Estrogen regulates expression of tumor necrosis factor receptors in breast adipose fibroblasts. J Clin Endocrinol Metab 89: 4018-4024. 187. Altucci L, Addeo R, Cicatiello L, Germano D, Pacilio C, et al. (1997) Estrogen induces early and timed activation of cyclin-dependent kinases 4, 5, and 6 and increases cyclin messenger ribonucleic acid expression in rat uterus. Endocrinology 138: 978-984. 188. Yokota T, Oritani K, Garrett KP, Kouro T, Nishida M, et al. (2008) Soluble frizzled- related protein 1 is estrogen inducible in bone marrow stromal cells and suppresses the earliest events in lymphopoiesis. J Immunol 181: 6061-6072. 189. Hill CS, Treisman R (1995) Transcriptional regulation by extracellular signals: mechanisms and specificity. Cell 80: 199-211. 190. Groom LA, Sneddon AA, Alessi DR, Dowd S, Keyse SM (1996) Differential regulation of the MAP, SAP and RK/p38 kinases by Pyst1, a novel cytosolic dual- specificity phosphatase. EMBO J 15: 3621-3632. 191. Karpuzoglu-Sahin E, Hissong BD, Ansar Ahmed S (2001) Interferon-gamma levels are upregulated by 17-beta-estradiol and diethylstilbestrol. J Reprod Immunol 52: 113-127. 192. Lengi AJ, Phillips RA, Karpuzoglu E, Ahmed SA (2007) Estrogen selectively regulates chemokines in murine splenocytes. J Leukoc Biol 81: 1065-1074. 193. Guo TL, Chi RP, Germolec DR, White KL, Jr. (2005) Stimulation of the immune response in B6C3F1 mice by genistein is affected by exposure duration, gender, and litter order. J Nutr 135: 2449-2456.

166

194. Tiwari-Woodruff S, Morales LB, Lee R, Voskuhl RR (2007) Differential neuroprotective and antiinflammatory effects of estrogen receptor (ER)alpha and ERbeta ligand treatment. Proc Natl Acad Sci U S A 104: 14813-14818. 195. Nakaya M, Tachibana H, Yamada K (2006) Effect of estrogens on the interferon- gamma producing cell population of mouse splenocytes. Biosci Biotechnol Biochem 70: 47-53. 196. Mammen EF (2000) Oral contraceptive pills and hormonal replacement therapy and thromboembolic disease. Hematol Oncol Clin North Am 14: 1045-1059, vii-viii. 197. Mizejewski GJ, Dias JA, Hauer CR, Henrikson KP, Gierthy J (1996) Alpha- fetoprotein derived synthetic peptides: assay of an estrogen-modifying regulatory segment. Mol Cell Endocrinol 118: 15-23. 198. Ohashi R, Sugimura M, Kanayama N (2003) Estrogen administration enhances thrombin generation in rats. Thromb Res 112: 325-328. 199. Moro L, Reineri S, Piranda D, Pietrapiana D, Lova P, et al. (2005) Nongenomic effects of 17beta-estradiol in human platelets: potentiation of thrombin-induced aggregation through estrogen receptor beta and Src kinase. Blood 105: 115-121. 200. Zhang Y, Saha S, Rosenfeld G, Gonzalez J, Pepeljugoski KP, et al. (2010) Raloxifene modulates estrogen-mediated B cell autoreactivity in NZB/W F1 mice. J Rheumatol 37: 1646-1657. 201. Roselli M, Mineo TC, Basili S, Martini F, Mariotti S, et al. (2004) Soluble CD40 ligand plasma levels in lung cancer. Clin Cancer Res 10: 610-614. 202. Geraldes P, Gagnon S, Hadjadj S, Merhi Y, Sirois MG, et al. (2006) Estradiol blocks the induction of CD40 and CD40L expression on endothelial cells and prevents neutrophil adhesion: an ERalpha-mediated pathway. Cardiovasc Res 71: 566-573. 203. Hou H, Ge Z, Ying P, Dai J, Shi D, et al. (2012) Biomarkers of deep venous thrombosis. J Thromb Thrombolysis 34: 335-346. 204. Daly E, Vessey MP, Painter R, Hawkins MM (1996) Case-control study of venous thromboembolism risk in users of hormone replacement therapy. Lancet 348: 1027. 205. Vandenbroucke JP, Rosing J, Bloemenkamp KW, Middeldorp S, Helmerhorst FM, et al. (2001) Oral contraceptives and the risk of venous thrombosis. N Engl J Med 344: 1527-1535. 206. Rosendaal FR (2005) Venous thrombosis: the role of genes, environment, and behavior. Hematology Am Soc Hematol Educ Program: 1-12. 207. Goldhaber SZ (2005) Tamoxifen: preventing breast cancer and placing the risk of deep vein thrombosis in perspective. Circulation 111: 539-541. 208. Vandenberg LN, Hauser R, Marcus M, Olea N, Welshons WV (2007) Human exposure to bisphenol A (BPA). Reprod Toxicol 24: 139-177. 209. Zalko D, Jacques C, Duplan H, Bruel S, Perdu E (2011) Viable skin efficiently absorbs and metabolizes bisphenol A. Chemosphere 82: 424-430. 210. Lang IA, Galloway TS, Scarlett A, Henley WE, Depledge M, et al. (2008) Association of urinary bisphenol A concentration with medical disorders and laboratory abnormalities in adults. JAMA 300: 1303-1310.

167

211. Palanza P, Gioiosa L, vom Saal FS, Parmigiani S (2008) Effects of developmental exposure to bisphenol A on brain and behavior in mice. Environ Res 108: 150- 157. 212. Rubin BS, Lenkowski JR, Schaeberle CM, Vandenberg LN, Ronsheim PM, et al. (2006) Evidence of altered brain sexual differentiation in mice exposed perinatally to low, environmentally relevant levels of bisphenol A. Endocrinology 147: 3681-3691. 213. Honma S, Suzuki A, Buchanan DL, Katsu Y, Watanabe H, et al. (2002) Low dose effect of in utero exposure to bisphenol A and diethylstilbestrol on female mouse reproduction. Reprod Toxicol 16: 117-122. 214. Leranth C, Hajszan T, Szigeti-Buck K, Bober J, MacLusky NJ (2008) Bisphenol A prevents the synaptogenic response to estradiol in hippocampus and prefrontal cortex of ovariectomized nonhuman primates. Proc Natl Acad Sci U S A 105: 14187-14191. 215. Elsworth JD, Jentsch JD, Vandevoort CA, Roth RH, Jr DE, et al. (2013) Prenatal exposure to bisphenol A impacts midbrain dopamine neurons and hippocampal spine synapses in non-human primates. Neurotoxicology 35: 113-120. 216. Ho SM, Tang WY, Belmonte de Frausto J, Prins GS (2006) Developmental exposure to estradiol and bisphenol A increases susceptibility to prostate carcinogenesis and epigenetically regulates phosphodiesterase type 4 variant 4. Cancer Res 66: 5624-5632. 217. Markey CM, Luque EH, Munoz De Toro M, Sonnenschein C, Soto AM (2001) In utero exposure to bisphenol A alters the development and tissue organization of the mouse mammary gland. Biol Reprod 65: 1215-1223. 218. Munoz-de-Toro M, Markey CM, Wadia PR, Luque EH, Rubin BS, et al. (2005) Perinatal exposure to bisphenol-A alters peripubertal mammary gland development in mice. Endocrinology 146: 4138-4147. 219. Beronius A, Ruden C, Hakansson H, Hanberg A (2010) Risk to all or none? A comparative analysis of controversies in the health risk assessment of Bisphenol A. Reprod Toxicol 29: 132-146. 220. Calafat AM, Weuve J, Ye X, Jia LT, Hu H, et al. (2009) Exposure to bisphenol A and other phenols in neonatal intensive care unit premature infants. Environ Health Perspect 117: 639-644. 221. Poole A, van Herwijnen P, Weideli H, Thomas MC, Ransbotyn G, et al. (2004) Review of the toxicology, human exposure and safety assessment for bisphenol A diglycidylether (BADGE). Food Addit Contam 21: 905-919. 222. Fromme H, Kuchler T, Otto T, Pilz K, Muller J, et al. (2002) Occurrence of phthalates and bisphenol A and F in the environment. Water Res 36: 1429-1438. 223. Gallart-Ayala H, Moyano E, Galceran MT (2011) Analysis of bisphenols in soft drinks by on-line solid phase extraction fast liquid chromatography-tandem mass spectrometry. Anal Chim Acta 683: 227-233. 224. Liao C, Liu F, Guo Y, Moon HB, Nakata H, et al. (2012) Occurrence of eight bisphenol analogues in indoor dust from the United States and several Asian countries: implications for human exposure. Environ Sci Technol 46: 9138-9145.

168

225. Vinas P, Campillo N, Martinez-Castillo N, Hernandez-Cordoba M (2010) Comparison of two derivatization-based methods for solid-phase microextraction- gas chromatography-mass spectrometric determination of bisphenol A, bisphenol S and biphenol migrated from food cans. Anal Bioanal Chem 397: 115-125. 226. Grumetto L, Montesano D, Seccia S, Albrizio S, Barbato F (2008) Determination of bisphenol a and bisphenol B residues in canned peeled tomatoes by reversed- phase liquid chromatography. J Agric Food Chem 56: 10633-10637. 227. Cobellis L, Colacurci N, Trabucco E, Carpentiero C, Grumetto L (2009) Measurement of bisphenol A and bisphenol B levels in human blood sera from healthy and endometriotic women. Biomed Chromatogr 23: 1186-1190. 228. Matthews JB, Twomey K, Zacharewski TR (2001) In vitro and in vivo interactions of bisphenol A and its metabolite, bisphenol A glucuronide, with estrogen receptors alpha and beta. Chem Res Toxicol 14: 149-157. 229. Kitamura S, Suzuki T, Sanoh S, Kohta R, Jinno N, et al. (2005) Comparative study of the endocrine-disrupting activity of bisphenol A and 19 related compounds. Toxicol Sci 84: 249-259. 230. Matsushima A, Liu X, Okada H, Shimohigashi M, Shimohigashi Y (2010) Bisphenol AF is a full agonist for the estrogen receptor ERalpha but a highly specific antagonist for ERbeta. Environ Health Perspect 118: 1267-1272. 231. Satoh K, Ohyama K, Aoki N, Iida M, Nagai F (2004) Study on anti-androgenic effects of bisphenol a diglycidyl ether (BADGE), bisphenol F diglycidyl ether (BFDGE) and their derivatives using cells stably transfected with human androgen receptor, AR-EcoScreen. Food Chem Toxicol 42: 983-993. 232. Nakazawa H, Yamaguchi A, Inoue K, Yamazaki T, Kato K, et al. (2002) In vitro assay of hydrolysis and chlorohydroxy derivatives of bisphenol A diglycidyl ether for estrogenic activity. Food Chem Toxicol 40: 1827-1832. 233. Delfosse V, Grimaldi M, Pons JL, Boulahtouf A, le Maire A, et al. (2012) Structural and mechanistic insights into bisphenols action provide guidelines for risk assessment and discovery of bisphenol A substitutes. Proc Natl Acad Sci U S A 109: 14930-14935. 234. Cosnefroy A, Brion F, Maillot-Marechal E, Porcher JM, Pakdel F, et al. (2012) Selective activation of zebrafish estrogen receptor subtypes by chemicals by using stable reporter gene assay developed in a zebrafish liver cell line. Toxicol Sci 125: 439-449. 235. Van den Belt K, Verheyen R, Witters H (2003) Comparison of vitellogenin responses in zebrafish and rainbow trout following exposure to environmental estrogens. Ecotoxicol Environ Saf 56: 271-281. 236. Wang J, Liu X, Wang H, Wu T, Hu X, et al. (2010) Expression of two cytochrome P450 aromatase genes is regulated by endocrine disrupting chemicals in rare minnow Gobiocypris rarus juveniles. Comp Biochem Physiol C Toxicol Pharmacol 152: 313-320. 237. Schafers C, Teigeler M, Wenzel A, Maack G, Fenske M, et al. (2007) Concentration- and time-dependent effects of the synthetic estrogen, 17alpha-

169

ethinylestradiol, on reproductive capabilities of the zebrafish, Danio rerio. J Toxicol Environ Health A 70: 768-779. 238. Perez P, Pulgar R, Olea-Serrano F, Villalobos M, Rivas A, et al. (1998) The estrogenicity of bisphenol A-related diphenylalkanes with various substituents at the central carbon and the hydroxy groups. Environ Health Perspect 106: 167-174. 239. Chen MY, Ike M, Fujita M (2002) Acute toxicity, mutagenicity, and estrogenicity of bisphenol-A and other bisphenols. Environ Toxicol 17: 80-86. 240. Grignard E, Lapenna S, Bremer S (2012) Weak estrogenic transcriptional activities of Bisphenol A and Bisphenol S. Toxicol In Vitro 26: 727-731. 241. Kuruto-Niwa R, Nozawa R, Miyakoshi T, Shiozawa T, Terao Y (2005) Estrogenic activity of alkylphenols, bisphenol S, and their chlorinated derivatives using a GFP expression system. Environ Toxicol Pharmacol 19: 121-130. 242. Hashimoto Y, Nakamura M (2000) Estrogenic activity of dental materials and bisphenol-A related chemicals in vitro. Dent Mater J 19: 245-262. 243. Hashimoto Y, Moriguchi Y, Oshima H, Kawaguchi M, Miyazaki K, et al. (2001) Measurement of estrogenic activity of chemicals for the development of new dental polymers. Toxicol In Vitro 15: 421-425. 244. Olea N, Pulgar R, Perez P, Olea-Serrano F, Rivas A, et al. (1996) Estrogenicity of resin-based composites and sealants used in dentistry. Environ Health Perspect 104: 298-305. 245. Terasaki M, Kazama T, Shiraishi F, Makino M (2006) Identification and estrogenic characterization of impurities in commercial bisphenol A diglycidyl ether (BADGE). Chemosphere 65: 873-880. 246. Reid EE WE (1944) The relation of estrogenic activity to structure in some 4,4'- dihydroxydephenylmethanes. J Am Chem Soc 66: 967-968. 247. Dodds EC LW (1936) Synthetic estrogenic agents without the phenanthrene nucleus. Nature 137: 996. 248. Kanai H, Barrett JC, Metzler M, Tsutsui T (2001) Cell-transforming activity and estrogenicity of bisphenol-A and 4 of its analogs in mammalian cells. Int J Cancer 93: 20-25. 249. Judson RS, Houck KA, Kavlock RJ, Knudsen TB, Martin MT, et al. (2010) In vitro screening of environmental chemicals for targeted testing prioritization: the ToxCast project. Environ Health Perspect 118: 485-492. 250. Reif DM, Martin MT, Tan SW, Houck KA, Judson RS, et al. (2010) Endocrine profiling and prioritization of environmental chemicals using ToxCast data. Environ Health Perspect 118: 1714-1720. 251. Eimon PM, Rubinstein AL (2009) The use of in vivo zebrafish assays in drug toxicity screening. Expert Opin Drug Metab Toxicol 5: 393-401. 252. Walker SL, Ariga J, Mathias JR, Coothankandaswamy V, Xie X, et al. (2012) Automated reporter quantification in vivo: high-throughput screening method for reporter-based assays in zebrafish. PLoS One 7: e29916. 253. Saydmohammed M, Vollmer LL, Onuoha EO, Vogt A, Tsang M (2011) A high- content screening assay in transgenic zebrafish identifies two novel activators of fgf signaling. Birth Defects Res C Embryo Today 93: 281-287.

170

254. Rocke J, Lees J, Packham I, Chico T (2009) The zebrafish as a novel tool for cardiovascular drug discovery. Recent Pat Cardiovasc Drug Discov 4: 1-5. 255. Legler JB, J.L.M, Brouwer, A., Lanser, P.H., Murk, A. J., vander Saag, P.T. et al., (2000) A novel in vivo bioassay for (xeno)-estrogens using transgenic zebrafish. . Eniron Sci Technol 34: 4439-4444. 256. Shelby MD, Newbold RR, Tully DB, Chae K, Davis VL (1996) Assessing environmental chemicals for estrogenicity using a combination of in vitro and in vivo assays. Environ Health Perspect 104: 1296-1300. 257. Elsby R, Maggs JL, Ashby J, Paton D, Sumpter JP, et al. (2001) Assessment of the effects of metabolism on the estrogenic activity of xenoestrogens: a two-stage approach coupling human liver microsomes and a yeast estrogenicity assay. J Pharmacol Exp Ther 296: 329-337. 258. Beresford N, Routledge EJ, Harris CA, Sumpter JP (2000) Issues arising when interpreting results from an in vitro assay for estrogenic activity. Toxicol Appl Pharmacol 162: 22-33. 259. Gaido KW, Maness SC, McDonnell DP, Dehal SS, Kupfer D, et al. (2000) Interaction of methoxychlor and related compounds with estrogen receptor alpha and beta, and androgen receptor: structure-activity studies. Mol Pharmacol 58: 852-858. 260. Orton F, Rosivatz E, Scholze M, Kortenkamp A (2011) Widely used pesticides with previously unknown endocrine activity revealed as in vitro antiandrogens. Environ Health Perspect 119: 794-800. 261. Teng Y, Manavalan TT, Hu C, Medjakovic S, Jungbauer A, et al. (2012) Endocrine disruptors fludioxonil and fenhexamid stimulate miR-21 expression in breast cancer cells. Toxicol Sci 131: 71-83. 262. Fillinger S, Leroux P, Auclair C, Barreau C, Al Hajj C, et al. (2008) Genetic analysis of fenhexamid-resistant field isolates of the phytopathogenic fungus Botrytis cinerea. Antimicrob Agents Chemother 52: 3933-3940. 263. Martin MT, Knudsen TB, Reif DM, Houck KA, Judson RS, et al. (2011) Predictive model of rat reproductive toxicity from ToxCast high throughput screening. Biol Reprod 85: 327-339. 264. Wade MG, Desaulniers D, Leingartner K, Foster WG (1997) Interactions between endosulfan and dieldrin on estrogen-mediated processes in vitro and in vivo. Reprod Toxicol 11: 791-798. 265. Chow WS, Chan WK, Chan KM (2012) Toxicity assessment and vitellogenin expression in zebrafish (Danio rerio) embryos and larvae acutely exposed to bisphenol A, endosulfan, heptachlor, methoxychlor and tetrabromobisphenol A. J Appl Toxicol 33: 670-678. 266. Oh YJ, Jung YJ, Kang JW, Yoo YS (2007) Investigation of the estrogenic activities of pesticides from Pal-dang reservoir by in vitro assay. Sci Total Environ 388: 8- 15. 267. Lemaire G, Mnif W, Mauvais P, Balaguer P, Rahmani R (2006) Activation of alpha- and beta-estrogen receptors by persistent pesticides in reporter cell lines. Life Sci 79: 1160-1169.

171

268. Matthews J, Wihlen B, Thomsen J, Gustafsson JA (2005) Aryl hydrocarbon receptor-mediated transcription: ligand-dependent recruitment of estrogen receptor alpha to 2,3,7,8-tetrachlorodibenzo-p-dioxin-responsive promoters. Mol Cell Biol 25: 5317-5328. 269. Handa RJ, Pak TR, Kudwa AE, Lund TD, Hinds L (2008) An alternate pathway for androgen regulation of brain function: activation of estrogen receptor beta by the metabolite of dihydrotestosterone, 5alpha-androstane-3beta,17beta-diol. Horm Behav 53: 741-752. 270. Jefferson WN, Padilla-Banks E, Clark G, Newbold RR (2002) Assessing estrogenic activity of phytochemicals using transcriptional activation and immature mouse uterotrophic responses. J Chromatogr B Analyt Technol Biomed Life Sci 777: 179-189. 271. Fremont L (2000) Biological effects of resveratrol. Life Sci 66: 663-673. 272. Ye HY, Lou YJ (2005) Estrogenic effects of two derivatives of icariin on human breast cancer MCF-7 cells. Phytomedicine 12: 735-741. 273. Liu J, Ye H, Lou Y (2005) Determination of rat urinary metabolites of icariin in vivo and estrogenic activities of its metabolites on MCF-7 cells. Pharmazie 60: 120-125. 274. Wang ZQ, Lou YJ (2004) Proliferation-stimulating effects of icaritin and desmethylicaritin in MCF-7 cells. Eur J Pharmacol 504: 147-153. 275. Joffe H, Cohen LS (1998) Estrogen, serotonin, and mood disturbance: where is the therapeutic bridge? Biol Psychiatry 44: 798-811. 276. Ryan J, Ancelin ML (2012) Polymorphisms of estrogen receptors and risk of depression: therapeutic implications. Drugs 72: 1725-1738. 277. Wilson CA, Hunter AJ (1985) Progesterone stimulates sexual behaviour in female rats by increasing 5-HT activity on 5-HT2 receptors. Brain Res 333: 223-229. 278. Ou HC, Santos F, Raible DW, Simon JA, Rubel EW (2010) Drug screening for hearing loss: using the zebrafish lateral line to screen for drugs that prevent and cause hearing loss. Drug Discov Today 15: 265-271.

172