Mechanisms of Fgf8 Transcription in the Developing Mouse

MECHANISMS OF FGF8 TRANSCRIPTION IN THE DEVELOPING MOUSE

OLFACTORY PLACODE

A dissertation submitted to

Kent State University in partial

fulfillment of the requirements for the

degree of Doctor of Philosophy

Megan L. Linscott

May, 2020

Except for previously published materials

Dissertation written by

Megan L. Linscott

B.S., Kent State University, 2015

Ph.D., Kent State University, 2020

Approved by

Wilson C.J. Chung, Ph.D. , Chair, Doctoral Dissertation Committee

Kristy Welshhans, Ph.D., Members, Doctoral Dissertation Committee

Jennifer McDonough, Ph.D.

Bansidhar Datta, Ph.D.

Mary Ann Raghanti, Ph.D.

Accepted by

Laura G. Leff, Ph.D. , Chair, Department of Biological Sciences

James L. Blank, Ph.D. , Dean, College of Arts and Sciences TABLE OF CONTENTS TABLE OF CONTENTS ...... iii LIST OF FIGURES ...... v LIST OF TABLES ...... vii LIST OF ABBREVIATIONS……………………………………………………………….….viii ACKNOWLEDGEMENTS ...... xi

I. GENERAL INTRODUCTION...... 1

REFERENCES………………………………………………………….……….20

II. FIBROBLAST GROWTH FACTOR 8 EXPRESSION IN GT1-7 GNRH- SECRETING NEURONS IS ANDROGEN-INDEPENDENT, BUT CAN BE UPREGULATED BY THE INHIBITION OF DNA METHYLTRANSFERASES………………………...... 39

INTRODUCTION...... 39

MATERIALS AND METHODS...... 41

RESULTS...... 49

DISCUSSION...... 57

REFERENCES...... 61

III. THE INDIRECT ROLE OF RETINOIC ACID IN THE DEVELOPING OLFACTORY PLACODE………………………………………………………66

INTRODUCTION...... 66

MATERIALS AND METHODS...... 68

RESULTS...... 72

DISCUSSION...... 85

REFERENCES...... 88

iii

IV. TET1 REGULATES FIBROBLAST GROWTH FACTOR 8 TRANSCRIPTION IN GONADOTROPIN RELEASING HORMONE NEURONS ………..……...92

INTRODUCTION...... 92

MATERIALS AND METHODS...... 96

RESULTS...... 104

DISCUSSION...... 118

SUPPORTING INFORMATION…………………..…..………………………124

REFERENCES...... 129

V. TET1 NULL MUTATION IN MICE DISRUPTS EMBRYONIC DEVELOPMENT OF GNRH NEURONS AND THEIR PHYSIOLOGICAL FUNCTION IN ADULTHOOD …………………………...…………………..139

INTRODUCTION...... 139

MATERIALS AND METHODS...... 141

RESULTS...... 144

DISCUSSION...... 152

REFERENCES...... 155

VI. GENERAL DISCUSSION ………………………………..………….………..159

REFERENCES……………………………………………..…………………..171

LIST OF FIGURES

Figure 2.1. Immunocytochemistry of GnRH expressing GT1-7 neurons………………………..43

Figure 2.2. ChIP gel electrophoresis in vehicle or DHT treated GT1-7 neurons and RT-PCR analysis of ARE 1-3 sites...... 50

Figure 2.3. RT-qPCR for Fgf8 and Fgfr1 mRNA in GT1-7 neurons treated for 2 hrs with vehicle or 100 nM DHT, 4 hrs with vehicle or 100 nM DHT, or 48 hrs with vehicle or 100 nM

DHT...…………………...... 52

Figure 2.4. RT-qPCR for Fgf8, Fgfr1, GnRH, AR mRNA nasal explants treated for 3 DIV with vehicle or 100 nM DHT.……………………………………………………...... 54

Figure 2.5. RT-qPCR for Fgf8 and GnRH mRNA in GT1-7 neurons treated with vehicle or 1 µM

AZA for 72 hours...... 56

Figure 3.1. RALDH expression in the OP...... 74

Figure 3.2. Fgf8 expression in the GN11 cell line...... 76

Figure 3.3. RA represses GnRH expression in the OP...... 78

Figure 3.4. RARs do not bind to the Fgf8 promoter during OP development…………………...80

Figure 3.5. RALDH inhibition induces Fgf8 expression in the OP……………………….……..82

Figure 3.6. ATRA enhances GN11 neuron migration and/or proliferation………………….…..84

Figure 4.1. Schematic of transient Fgf8 transcription during GnRH neuronal emergence in the embryonic mouse OP...... 94

Figure 4.2. AZA induced Fgf8 expression in the OP...... 105

Figure 4.3. TET or DNMT expression in the OP...... 107

Figure 4.4. DNMT3b binds to the Fgf8 promoter early in OP development…………………...109

Figure 4.5. 5hmc accumulation on the Fgf8 promoter is driven in a time-dependent fashion…………………………………………………………………………………………..111

Figure 4.6. Epigenetic switch on the Fgf8 promoter in the E9.5 and E13.5 OP...... 113

Figure 4.7. Fgf8 histone modifications during GnRH neuron ontogenesis...... 115

Figure 4.8. Tet1 siRNA knockdown in GT1-7 neurons……………………………....………...117

Figure 4.9. Schematic model of age dependent Fgf8 transcriptional control in the embryonic mouse OP……………………………………………………...……………………….……….123

Figure S4.10. Negative MeDIP qPCR in 3-5 pooled mouse OPs at E9.5 on the 3’UTR

Fgf8……………………………………………………………………………...……...………124

Figure S4.11. Propidium iodine staining in GT1-7 cells treated with AZA for 3 days in vitro……………………………………………………………………………………………..125

Figure S4.12. Significant levels of 5mc on the Fgf8 promoter at E8.5………….……………..126

Figure S4.13. Fgf8 promoter landscape in the E10.5 midbrain-hindbrain…………..……..…..127

Figure S4.14. EZH2 is highly expressed in the developing OP……...……………………..…..128

Figure 5.1. Heat map of the E13.5 mouse OP in WT versus TET1-/-...... 145

Figure 5.2. Reproductive capacity of TET1-/- mice………...………………...... 147

Figure 5.3. Immunoreactivity of GnRH in female WT or TET1-/- PN0 mice…………..………149

Figure 5.4. Fgf8, Fgfr1, and GnRH mRNA expression in the E13.5 OP of WT versus TET1-/- animals……………………………………………..…………………………………..……….151

Figure 6.1. Schematic of Fgf8 activation during GnRH neuron ontogenesis...... 166

LIST OF TABLES

Table 2.1. Primer sequences used for detecting Fgf8, Fgfr1, GnRH and AR mRNA expression, and ARE containing DNA sequences within the Fgf8 promoter region……………...... 46

Table 4.1. Forward and reverse primer sequences for detection of genomic DNA and mRNA transcripts used in this study...... 99

vii

LIST OF ABBREVIATIONS

ACC Agenesis of the corpus callosum

ANOVA Analysis of variance

AR Androgen receptor

ATRA All-trans-retinoic acid

AZA 5-Azacytidine

CHH Congenital hypogonadotropic

ChIP hypogonadism

Chromatin immunoprecipitation

CpG 5'—Cytosine—phosphate—

Guanine—3'

CRH Corticotropin releasing hormone

DEXA Dual X-ray absorptiometry

DNMT DNA methyltransferase

EZH Enhancer of zeste

E Embryonic day

FGF Fibroblast growth factor

FGFR Fibroblast growth factor receptor

FSH Follicle stimulating hormone

GnRH Gonadotropin releasing hormone

Hprt Hypoxanthine-guanine

phosphoribosyltransferase

viii

IHC Immunohistochemistry

IR Immunoreactivity

KP Kisspeptin

KS Kallmann Syndrome

LH Luteinizing hormone

MBHB Midbrain hindbrain

MeDIP Methylated immunoprecipitation

NP-1 Neuropilin-1

NSC Neural stem cell

OP Olfactory placode

PK Proteinase K

PN0 Postnatal day

POA Preoptic nucleus

PRC2 Polycomb repressive complex 2

PVN Paraventricular nucleus

RA Retinoic acid

RALDH Retinaldehyde dehydrogenase

RAR Retinoic acid receptor

RT-qPCR Real-time quantitative

polymerase chain reaction

SEM Standard error of the mean siRNA

TET Small interfering ribonucleic

acid

Ten-eleven translocation

TF Transcription factor

UTR Untranslated region

UV Ultraviolet

VP Vasopressin

WNT Wingless-related integration site

WT Wildtype

ACKNOWLEDGEMENTS

The completion of my PhD would not have been possible without the help and guidance of many friends and family. The overwhelming support I have received through this time period has been tremendous. I am so fortunate and lucky to have so many people in my life who push me to achieve my goals. With that, I would like to extend my most sincere thank you to everyone who has been there for me during these trying and stressful times.

To my committee: first and foremost, my mentor, Dr. Wilson Chung. I cannot thank you enough for putting up with me for all these years. You have provided me with so many opportunities to succeed and pushed me to be the scientist that I am today, which I at many points, did not think was possible. You have believed in me since day one. I could not have asked for a better mentor. To Dr. Kristy Welshhans, Dr. Jennifer McDonough and Dr. Bansidhar

Datta and, thank you all for your input and comments during my graduate studies.

To my lab mates: Courtney, you have mentored me since day one and taught me so much along the way. I am so lucky to have someone as bright as you to listen to my ideas. Thank you for covering me when experiments got crazy. I will never forget all the great food and fun we had at the many conferences we travelled to. I hope to see you in the future at a mutual conference! To Kathy and Anastasia, you are the funniest, most hardworking people I have ever met. It was a pleasure to mentor you. You made some of the most monotonous days brighter and weirder (which I mean in the best sense possible).

Finally, to family and friends: My parents, Dave and Tracy, you have raised me to be tenacious and self-sufficient. You have taught me to never give up and have believed in me without question every step of the way. To Sohini, I cannot imagine where I would be without your support, friendship and delicious curry.

CHAPTER I GENERAL INTRODUCTION

GnRH regulation of fertility

Human infertility is defined as the inability to conceive after one year of regular unprotected sex ("Report on optimal evaluation of the infertile male" 2006). Infertility arises in both the male and female reproductive systems and includes dysfunction of ovulation, spermatogenesis, uterine abnormalities, testicular atrophy and cervical factors.

Surprisingly, the American Society for Reproductive medicine indicates that approximately 15% of healthy individuals are infertile ("Report on optimal evaluation of the infertile male" 2006). Furthermore, according to the WHO, infertility and subinfertility afflicts approximately 10% of the world’s women who have been in a stable relationship for five to more years (Mascarenhas et al., 2012). The prevalence increases

2.5 times when also considering women using a two year conception time frame

(Mascarenhas et al., 2012). Taken together, these data lead to the conclusion that infertility or subfertility is a significant health burden.

Mammalian reproductive success depends on a small population of gonadotropin releasing hormones (GnRH) neurons to stimulate gonadotropin secretion from the anterior pituitary and activate gonadal steroidogenesis and gametogenesis. GnRH neuron cell bodies reside in the medial pre-optic area (POA) of the hypothalamus and send projection fibers to the median eminence where they secrete GnRH peptide into the portal

1 vein system of the anterior pituitary to cause LH and FSH release from the gonadotrophs, which then travels through the general circulation to the gonads which stimulates gametogenesis and the production of sex steroids.

Kallmann syndrome

Kallmann syndrome (KS) is a form of CHH that is associated with hyposmia or anosmia (Boehm et al., 2015; Chung et al., 2016; Cole LW, 2007; Pitteloud, 2007; P. S. Tsai et al., 2011). Most abnormalities result from early human embryonic development, between four and ten weeks of pregnancy. Generally, KS patients do not undergo puberty, and are infertile in adulthood.

Diagnosis typically arises from hyposmia or anosmia, which can be further confirmed with MRIs to detect the bilateral absence of olfactory bulbs. Furthermore, laboratory testing can identify deviations in hormone levels, as well as genetic screenings for known mutations associated with

KS. Many non-endocrine dysfunctions have also been reported for KS including mirror movements (Conrad et al., 1978; Quinton et al., 1996; Quinton et al., 2001; Schwanzel-Fukuda et al., 1989), corpus callosum agenesis (Dode et al., 2007; Dode et al., 2003), congenital ptosis

(J. Hardelin et al., 1993; Reardon, 2007), renal agenesis (J.-P. Hardelin et al., 1993; Wegenke et al., 1975), and palate abnormalities (Mølsted et al., 1997; Sarfati et al., 2010). Genetic screenings of individuals with GnRH deficiencies identified several genes important for GnRH neuron fate-specification, migration, homeostasis, and peptide release. Recent studies in Finland have estimated an overall 1 in 48,000 case with 1 in 30,000 in males and 1 in 125,000 in females

(Laitinen et al., 2011). X-linked inheritance is believed to be causal for a higher male prevalence, as anosmin-1, located on the X chromosome, is a common KS mutation (Franco et al., 1991;

Hardelin et al., 1992; Legouis et al., 1991).

Diagnosis is typically made during puberty, when sexual maturation is delayed or absent.

Treatment is primarily focused on hormonal replacement therapy, usually testosterone for males and combined estrogen/progesterone for women, which can induce secondary sex characteristics.

This can be combined with gonadotropin or pulsatile GnRH therapy to induce fertility.

Currently, both treatment options require excessive patient monitoring, as overstimulation of gonadal hormone systems can cause hypoestrogenism, hot flashes, osteoporosis, and insomnia.

Typical therapies continue short term (6 months), while long term therapy is often coupled with

“add-back” therapies for pain, estrogen, calcium and vitamin D, as well as regular DEXA scans for bone-mineral density.

Kallmann syndrome is heterogenous, with several known genes affecting all stages of

GnRH neuron development. The most common mutations are, anosmin-1 (Franco et al., 1991;

Hardelin et al., 1992; Legouis et al., 1991), fibroblast growth factor 8 (Fgf8) (W. C. Chung et al., 2008; Falardeau et al., 2008), fibroblast growth factor 17 (Fgf17) (Miraoui et al., 2013), fibroblast growth factor receptor-1 (Fgfr1) (Dode et al., 2003), interleukin-17 receptor D

(Miraoui et al., 2013), dual specificity phosphatase 6 (Miraoui et al., 2013), sprouty homolog 4

(Miraoui et al., 2013), fibronectin leucine rich transmembrane protein 3 (Miraoui et al., 2013),

β‐Klotho (Xu et al., 2017), prokineticin 2 (Dode et al., 2006), prokineticin receptor 2 (Cole et al., 2008), and chromodomain helicase-DNA-binding 7 (Kim et al., 2008). Despite this, reduced

GnRH function in patients harboring these mutations only accounts for about 30 percent of KS cases, indicating a role for environmental factors or non-genetic abnormalities.

Further analyses found that these gene mutations oftentimes cause abnormal embryonic

GnRH neuron development, the primary reason as to why GnRH signaling is absent in KS patients. For example, a groundbreaking study in a human embryo with an inherited form of KS

3 found evidence that GnRH neuron migration was impaired. Indeed, GnRH neurons were detected in the cribriform plate rather than their normal location in the preoptic and hypothalamic region (Schwanzel-Fukuda et al., 1989). Later studies in humans and rodents support that defects in GnRH neuron migration is a major cause of KS (Cho et al., 2019; Young et al., 2019).

Nevertheless, ours and other studies indicate that defective GnRH neuron migration is not the only developmental cellular cause of KS. Alternatively, KS gene mutations may disrupt GnRH neuron ontogenesis and fate-specification in the olfactory placode (OP) (Chung et al., 2016;

Chung et al., 2010; W. C. Chung et al., 2008; Chung & Tsai, 2010; Cole LW, 2007; Falardeau et al., 2008; Forni et al., 2013; Kawauchi et al., 2007; Linscott & Chung, 2019; Rochester et al.,

2012; Sabado et al., 2012; P. S. Tsai et al., 2011; Tsai & Gill, 2006; Tsai et al., 2005).

Development of the olfactory placode

The OP, an ectodermal region that gives rise to (non-sensory) respiratory epithelium and

(sensory) olfactory epithelium (Forni et al., 2013; Kramer & Wray, 2000, 2001), contains a progenitor cell population of GnRH neurons (Schwanzel-Fukuda & Pfaff, 1989; Wray, Nieburgs, et al., 1989). In amphibians, removal of the OP results in the absence of olfactory epithelium, nerve, and bulb, as well as GnRH neurons of the forebrain (Murakami et al., 1992; Northcutt &

Muske, 1994). In mouse OP studies, a very small percentage of GnRH neurons were found to originate from the overlapping respiratory epithelium (Kramer & Wray, 2000). Additionally, migratory neural crest cells also contribute (~30%) to the GnRH progenitor pool, as they border the neural plate and overlap with the presumptive OP (Forni et al., 2011; Katoh et al., 2011;

Whitlock & Westerfield, 2000). The remaining 70% of the total population of GnRH neurons

4 were confirmed to be of placodal origin. Overall, mammalian GnRH neurons originate from both neural crest and OP cells.

About 80% of GnRH neurons become post-mitotic in the medial OP (mOP) around E9.5 or 10.5 (Wray, Grant, et al., 1989), indicating GnRH neurons are likely fate-specified in the mOP before their emergence. Additionally, one study found that approximately 1900-2000 embryonic GnRH neurons could be detected at E12.75, suggesting a selection process after

GnRH neuron differentiation (Wu et al., 1997). GnRH neurons then migrate caudally into the hypothalamus along vomeronasal nerves, where they extend their processes to the median eminence, and secrete GnRH into the capillaries of the median eminence to modulate LH and

FSH secretion. While dysregulation of GnRH neuron migration is commonly associated with

KS, the migratory process that regulates GnRH neurons is not well-understood. Indeed, in KS, mutations in axon guidance molecules (Luca & Comoglio Paolo, 2004), such as Sema3a, may be causal for axon guidance defects.

The role of FGF8 in GnRH neuron development

Fibroblast growth factor signaling is required for the embryonic development of the GnRH neuron system in humans and rodents (Chung et al., 2010; W. C. Chung et al., 2008; Falardeau et al., 2008; J. C. Gill et al., 2004; J. C. Gill & P. S. Tsai, 2006; Sidhoum et al., 2014; Tsai et al.,

2005). Furthermore, FGF8 has been shown to regulate fate specification, proliferation and differentiation of neuroendocrine cells. In postnatal animals, FGF8 signaling can also affect the oxytocin, vasopressin, kisspeptin and corticotropin releasing hormone neurons in the hypothalamus (Brooks et al., 2010; Brooks et al., 2012; Rodriguez et al., 2015; Tata et al.,

2012). In humans, mutations in FGFR1 and/or FGF8 cause KS symptoms, such as absence of

GnRH release, non-detectable levels of LH and FSH, absence of pubertal onset to ultimately cause infertility in adulthood (Albuisson et al., 2005; Cole LW, 2007; Dode et al., 2003;

Falardeau et al., 2008). Basic research studies in mice carrying a dominant-negative form of

FGFR1 found that adult mice have a reduced number of GnRH neurons and undergo early reproductive senescence (J. C. Gill et al., 2004; J. C. Gill & P. S. Tsai, 2006). Similarly, newborn

Fgfr1 hypomorphic mice exhibited a ~80% reduction in the total number of GnRH neurons in the entire head (W. C. Chung et al., 2008). Further analyses demonstrated that the dramatic decrease in GnRH neuron number was not due to obvious migration defects, since no abnormal accumulation of GnRH neurons along the GnRH neuron migration route from the OP to the preoptic/hypothalamic region were detected (W. C. Chung et al., 2008).

Reduced FGF8 signaling in mice, showed an even larger deleterious effect on the developing GnRH neuron system. Indeed, embryonic (i.e., E11.5 and E14.5) and newborn homozygous Fgf8 hypomorphic mice did not harbor any GnRH neurons (W. C. Chung et al.,

2008; Falardeau et al., 2008). These data suggest that reduced FGF8 signaling in Fgf8 hypomorphic mice caused GnRH neurons not to emerge from their primary birthplace, the OP.

In contrast, GnRH neurons were detectable in heterozygous Fgf8 hypomorphs, albeit significantly reduced when compared to wildtype littermates (W. C. Chung et al., 2008;

Falardeau et al., 2008). Like Fgfr1 hypomorphic mice, no abnormally localized GnRH neurons were detected along the GnRH neuron migration pathway from the OP to the preoptic/hypothalamic region. Taken together, these results indicate that FGF-dependent loss of

GnRH neurons likely occurred during the emergence phase of GnRH neuron development prior to their migration (W. C. Chung et al., 2008).

Further developmental studies found that the GnRH neuron system in mice with heterozygous hypomorphy for Fgf8 and Fgfr1 was more disrupted than in mice with single heterozygous Fgf8 or Fgfr1 hypomorphy (Chung et al., 2010). In contrast, mice with digenic

Fgf8 and Fgfr3 defects showed no compounding deleterious effects in the GnRH neuron system

(Chung et al., 2010). Although studies found developing GnRH neurons to express FGFR1 and

FGFR3, (W. C. Chung et al., 2008; J. C. Gill et al., 2004; J. C. Gill & P. S. Tsai, 2006) FGF8 is able to signal through both receptor types (Ornitz & Itoh, 2015; Ornitz et al., 1996; Zhang et al.,

2006). Our observations indicate that FGF8-dependent GnRH neuron emergence from the mouse

OP is FGFR1-dependent rather than FGFR3. KS patients also harbor digenic loss-of-function mutations (Dode et al., 2006; Pitteloud et al., 2007). In contrast, mice with digenic Fgf8 and

Fgfr3 defects showed no compounding deleterious effects in the GnRH neuron system (Chung et al., 2010). These data also indicate that the worsened clinical reproductive outcomes in KS/CHH patients harboring digenic FGF8/FGFR1 loss-of-function mutations are likely the results on the compounding deleterious effects of Fgf8 and Fgfr1 mutations on GnRH neuron emergence

(Dode et al., 2006; Pitteloud et al., 2007).

Taken together, the data from these transgenic mouse studies clearly indicate that the absence of GnRH signaling in KS may not solely be due to abnormal migration as reported for inherited KS (Schwanzel-Fukuda et al., 1989), but can also be the result of failed GnRH neuron emergence prior to the migration phase. While FGF8 signaling controls cell proliferation and cell survival, we currently do not know whether the elimination of the GnRH progenitor cells is due to FGF8-dependent proliferation or cell survival. Because cell lineage studies confirmed that

Fgf8 mRNA is primarily localized in the respiratory epithelium (Bachler & Neubuser, 2001), it is likely that FGF8 expressing respiratory epithelial cells secrete FGF8 to provide trophic support

7 to promote cell proliferation or survival of GnRH (progenitor) cells that are primarily found in the olfactory epithelium of the OP. These studies, together with ones in Fgf8 and Fgfr1 hypomorphic mice and chicken (Sabado et al., 2012), demonstrated that the developmental timing of Fgf8 transcriptional activation and inhibition is critical during the GnRH neuronal emergence in the OP.

Transcriptional control of Fgf8

Classically, FGFs have been viewed as a potent morphogen, which instructs multiple downstream signaling cascades. The role of FGF8 signaling during embryo patterning is required for many aspects of development including body axis extension, limb development and brain patterning (Alexandre et al., 2006; Creuzet et al., 2004; Cunningham et al., 2015; Hoch et al.,

2015; Sirbu & Duester, 2006). Previous studies indicated that Fgf8 is expressed during a narrow window in the developing OP, which coincided with the period of GnRH neuron emergence (W.

C. Chung et al., 2008; Falardeau et al., 2008; Kawauchi et al., 2005; Sabado et al., 2012).

Previous studies indicated that control of Fgf8 transcription is dependent on transcription factors and epigenetic factors that control chromatin and/or histone modifications in human prostatic and breast cancer cells (Gnanapragasam et al., 2002; Kouhara et al., 1994; Saito et al., 1991; Tanaka et al., 1992; Yamanishi et al., 1995) or during body axis trunk development (Kumar & Duester,

2014). It is, however, unclear how Fgf8 transcription is initiated during OP embryogenesis.

Specifically, our goal is to understand how Fgf8 transcription is controlled during the period of

GnRH neuron emergence in mice. Recruitment of transcription factors or chromatin modifications are just two of the many possible mechanisms that could explain Fgf8 transcriptional regulation during this period of development.

Androgen receptor activation of Fgf8

FGF8 was originally discovered as an androgen-induced growth factor in SC-3 cells, an androgen-dependent mouse breast cancer cell line (Tanaka et al., 1992). Studies also showed that testosterone induced Fgf8 mRNA in 60-70% of newly diagnosed prostate cancers (Leung et al.,

1996). These results were recapitulated In LnCAP cells, a human prostate cancer cell line, which demonstrated that Fgf8 mRNA increased in the presence of androgen (Gnanapragasam et al.,

2002). Gnanapragasam et.al. specifically showed that sequences on the Fgf8 promoter could bind AR and activate Fgf8 activity. Similarly, site mutagenesis, as well as anti-androgen treatment of the three identified Androgen Response Elements (ARE) on the Fgf8 promoter could reverse Fgf8 promoter activation (Gnanapragasam et al., 2002). These collective studies underscore the ability of AR regulation of Fgf8 and suggest a direct activation mechanism.

Interestingly, during this time in development, the embryo has not undergone sexual differentiation, a process that occurs after GnRH neuron emerge.

Retinoic acid repression of Fgf8

RA signaling has been extensively evaluated in the developing embryo as a negative regulator of

Fgf8 transcription. For example, in the chick OP, ATRA-soaked beads reduced Fgf8 mRNA production and results in an elimination of GnRH neurons (Sabado et al., 2012). Additionally,

ATRA treatment in early GnRH neurons has been shown to reduce GnRH neuron emergence by not only reducing Fgf8 mRNA production but by direct binding events, as demonstrated by promoter activity assays (Cho et al., 1998; Sabado et al., 2012). While endogenous 9-cis RA is absent in the developing OP, ATRA is abundant and synthesized locally by RALDH3, which

9 activates RA receptors (RAR) α, β, ɣ in the OP (Denis et al., 2011; Forni & Wray, 2015; Sabado et al., 2012). RA can bind RARs or RXRs as homodimeric or heterodimeric complexes, respectively, to affect gene transcription. Importantly, recent studies in the developing embryonic trunk demonstrated the ability of RARs to bind a distal region on the Fgf8 promoter to exert a repressive effect (Kumar & Duester, 2014). This binding event also co-recruited members of the

PRC2 complex, which altered the chromatin structure of the Fgf8 promoter to a more repressive state (Kumar & Duester, 2014). These results suggest that Fgf8 expression in the OP may be influenced by local RA synthesis.

Epigenetics

Recent studies in the neuroendocrine system have focused on the field of epigenetics, which provided additional pathways that control gene expression other than transcription factors, co- factors or mutations (Aylwin et al., 2019; Chung & Auger, 2013; Gray et al., 2017). The definitions describing epigenetics ranges from those that include heritable gene function to those stating that epigenetic changes are molecular events that remodel chromatin without altering the primary genetic code (Bird, 2007). Epigenetic events on DNA or histones can be divided into two categories: repression or activation. DNA methylation initiates a cascade of events leading to gene repression and occurs when a methyl group is attached to cytosine within a 5’-CpN-3’ dinucleotide. This enzymatic reaction is catalyzed by DNA cytosine-5-methyltransferases

(DNMTs) (Grafstrom et al., 1985; Ramsahoye et al., 2000). The act of CpG site methylation typically does not have a direct impact on gene transcription rates, but rather, inhibition of gene transcription occurs when bound methyl-CpG-binding proteins recruit co-repressor complexes as opposed to co-activators. The histone deacetylase activity from these co-repressor complexes

10 will deacetylate histones, which contributes to chromatin condensation, and consequently gene repression (Bird & Wolffe, 1999; Klose & Bird, 2006; Yoon et al., 2003).

Demethylation of methylated CpG dinucleotides is catalyzed by the family of Ten-Eleven

Translocation (TET) proteins. Methylated cytosine residues can then be converted to 5- hydroxymethyl (5hmC), 5-formyl (5fC), and 5-carboxyl (5caC) (Wu & Zhang, 2011) by the TET proteins. Loss of TET proteins results in an overall decrease in gene expression and demethylation at promoters of active genes (Dawlaty et al., 2013; Ito et al., 2010). TET1 and

TET3 contain a CXXC domain, which directly binds to target DNA, and a C-terminal Fe+(II) and 2-ketoglutarate-dependent catalytic domain which progressively converts 5mC<5hmC<5fC<

5caC resulting in active or passive DNA demethylation. While TET3 binding patterns are found at diverse genomic elements, TET1 binding has been shown to be more exclusively associated near transcription start sites and at distal regulatory regions, presumably to affect long-range transcriptional interactions with for example, enhancer or silencer regulatory sequences (Huang et al., 2014). TET1 also can recruit many binding partners, including Enhancer Of Zeste 2

(EZH2; a histone 3 lysine 27 (H3K27) methyltransferase), which further complicates its regulatory role in DNA demethylation (Cartron et al., 2013; Wu et al., 2011; Wu & Zhang,

2011).

Previous studies have shown that DNA methylation is generally excluded from the TSS of TET1 bound, CpG-rich gene promoters, resulting in gene hypomethylation (Wu et al., 2011).

Moreover, loss of TET1 on CpG-rich promoters caused an increase in 5mC occupancy at TET1 bound loci (Wu et al., 2011). Interestingly, TET1-bound hypomethylated gene promoters have been associated with bivalent domains (Kong et al., 2016; Neri et al., 2013; Verma et al., 2018;

Voigt et al., 2013; Wu et al., 2011), (i.e., the presence of trimethylated histone 3 lysine 4

11 residues (H3K4me3) and H3K27me3 on the same or adjacent nucleosomes. Histone bivalency in embryonic development has proven to be an essential modification for genes responsible for cellular differentiation (Bernstein et al., 2006; Fisher & Fisher, 2011; Grandy et al., 2015).

Indeed, TET1 depletion decreases the binding of PRC2 to target genes, specifically by association with its binding partner EZH2 (Cartron et al., 2013).

Overall, these observations lead to the conclusion that TET1 has dual functions in gene regulation: DNA methylation and promoting histone bivalency through PRC2 recruitment.

Interestingly, studies showed that 5hmC is not merely a transitional chemical step during DNA demethylation. Indeed, 5hmC modifications are enriched in the central nervous system by up to ten times more than peripheral tissues (Branco et al., 2011); and moreover, differentiated neurons harbor more 5hmC than undifferentiated neurons (Maria A. Hahn et al., 2013). This also holds true after birth, where neuronal 5hmC levels rise, consistent with further maturation of neuronal processes, such as synaptogenesis, network integration, and proliferation (Münzel et al.,

2010). In 2010, Munzel et. al. demonstrated, using liquid chromatography and mass spectrometry of isotope labeled nucleosides, that levels of 5hmC are highest in the adult mammalian hypothalamus (Münzel et al., 2010). Therefore, the role of 5hmC and the development of neurons in the hypothalamus has been a major area of focus in our laboratory regarding GnRH neuron development.

Epigenetics in the HPG axis

Epigenetic factors represent non-mutational candidates underlying KS pathogenesis. For example, the GnRH promoter undergoes major chromatin and methylation changes (Iyer et al.,

2011). Additionally, only in mature GT1-7 cells did GnRH promoter elements display high

12 sensitivity to DNAseI treatment compared to non-neuronal cells and was only partially functional in immature GN11 cells (Iyer et al., 2011). These GnRH regulatory elements also displayed RNA polymerase II enrichment as well as the active histone marker, H3K4me3, indicating that maturation of GnRH neurons relies on chromatin changes on GnRH regulatory elements (Iyer et al., 2011). Indeed, one of the earliest studies of epigenetic regulation of GnRH neurons found that in Rhesus macaque monkeys, the GnRH promoter becomes demethylated at several, previously methylated CpG sites during GnRH neuronal development (Kurian &

Terasawa, 2013). Interestingly, GnRH gene demethylation was followed by a rise in GnRH mRNA expression, indicating that demethylation of the GnRH gene may regulate development and maturation of GnRH neurons. A similar demethylation mechanism operates during pubertal onset (Kurian & Terasawa, 2013). In vitro studies further found that the GnRH gene responded to external stimuli by altering chromatin modifications in mature GnRH neurons (Iyer et al.,

2011; Laverriere et al., 2016). For example, environmental pollutants, such as exposure to bisphenol A (BPA) readily altered the landscape of histone modifications at the GnRH promoter in mature GnRH neurons (Kurian et al., 2016). These results suggest that GnRH neuron development and pubertal onset require ongoing epigenetic modulation of genes.

Additionally, when peri-pubertal female rats were treated with the DNMT inhibitor azacitidine (AZA), they undergo delayed puberty, as measured by delayed vaginal opening and absence of estrous (Lomniczi et al., 2013; Lomniczi & Ojeda, 2016). Rats exposed to AZA had ovaries that were about half the size of control, pointing to demethylation as a contributor to abnormal ovarian follicle development (Lomniczi et al., 2013). Similarly, AZA exposure also disrupted kisspeptin (KP) function. Indeed, when medial basal hypothalamic fragments from postnatal day (P) 28 rats were exposed to KP, AZA-treated rats responded with significantly

13 more GnRH release than controls, pointing towards a deficit in KP production or other inputs to the GnRH system (Lomniczi et al., 2013). These results imply that development of the GnRH neuronal system during puberty is at least, partially under the influence of epigenetic processes.

However, a more comprehensive look at epigenetic sex-specific differences in early and pubertal hypothalamic development is required.

In females, TET1 knockout mice have abnormal ovarian follicle development with markedly smaller ovaries and few mature follicles (Dawlaty et al., 2013; Dawlaty et al., 2011).

Moreover, in 2016, Kurian et al. found that Tet2 mRNA was elevated in immature GN11 neurons versus adult GT1-7 neurons, and that Tet2 overexpression could increase GnRH mRNA and peptide by altering H3K4me3 abundance associated with the GnRH promoter (Kurian et al.,

2016). Moreover, GnRH-specific TET2 knockout mice displayed lower plasma LH levels and had lower fecundity in adult males, indicating the involvement of epigenetic factors during HPG development and function (Kurian et al., 2016). Interestingly, in vitro studies in gonadotroph cell lines (LβT2 and αT3-1 cells) demonstrated that TET1 can be directly repressed by GnRH and gonadal steroids. When treated with GnRH, both cell lines decreased Tet1 mRNA expression, and in αT3-1 cells treated with either estradiol or dihydrotestosterone, Tet1 mRNA decreased

(Yosefzon et al., 2017). Similarly, when intact animals were gonadectomized, Tet1 mRNA levels increased, while Tet2 and Tet3 mRNA were not affected (Yosefzon et al., 2017). Therefore,

TET1 is highly responsive to feedback from gonadotrophs in the pituitary, while TET2 effects

LH secretion from GnRH neurons.

Polycomb group proteins in the hypothalamus

Seminal studies in the hypothalamic arcuate nucleus showed that the Polycomb Group proteins

(PcG) control timing of puberty by repressing the Kiss1 gene. These results suggested that timing of puberty could be controlled by discrete chromatin modifiers. The polycomb group proteins are composed of four main subunits EZH1 or EZH2, suppressor of zeste (Suz12), and the WD40 domain proteins, EED and P55 that play a critical role in development, including differentiation and stem cell renewal (Chamberlain et al., 2008; Pasini et al., 2007). In neurons, deletion of

EZH2 enhanced neurogenesis of neural stem cells, and cortical-specific EZH2 deletion disrupted timing of cortical development (Hirabayashi et al., 2009; Pereira et al., 2010). The role of PcG proteins in the adult hypothalamic function, however, remains elusive. Nevertheless, groundbreaking studies found that in the female medial basolateral hypothalamus, decreased Eed and Cbx7 mRNA levels coincided with increased promoter methylation suggesting a role for

PcG proteins during the initiation of puberty (Lomniczi et al., 2013). Moreover, lentiviral delivery of EED to the median eminence-arcuate nucleus in P22 female rats inhibited GnRH pulse frequency, delayed puberty, disrupted estrous cyclicity, and impaired fertility (Lomniczi et al., 2013). In a recent study, pregnant mice exposed to BPA throughout pregnancy, caused delayed vaginal opening of female pups and an earlier first estrous. Methylation levels of Eed1,

Eed2 and Ezh2 from hypothalamic tissue were significantly higher in female offspring from BPA exposed groups (Su et al., 2019). While these studies underline an important role or PcG mediated pubertal activation, their function during GnRH neuron onset may also be of importance. Therefore, studies investigating the purpose of PcG and its components to control fate-specification during GnRH onset are critical to understand how mammals ensure both proper pubertal initiation and GnRH neuron development.

Chromatin and Epigenetic effects on Fgf8 transcription

While AR and RAR control of the Fgf8 promoter are driven by intrinsic sequences on the promoter region, other studies have demonstrated that Fgf8 can be activated through trans-acting regulatory features. Specifically, one study found that enhancers scattered along a 220 kb region could influence only Fgf8 activation (Marinic et al., 2013). Re-positioning of these sequences generated tissue specific Fgf8 expression in domains where Fgf8 is normally expressed (Marinic et al., 2013). These studies argue that the positioning of distal enhancers act as a unit to influence

Fgf8 gene expression, rather than cis sequences within the promoter, suggesting that the chromatin state of the Fgf8 locus, as well as insulating factors, must be receptive to modifications which arrange distal enhancers in correct positioning. Indeed, the Fgf8 chromatin state in non-expressing tissue has been shown to harbor a repressive H3K27me3 mark (Shen et al., 2012), however further investigation as to tissue specific activity/repression must be systematically evaluated to determine chromatin states. Additionally, multiple CpG islands upstream and downstream of the transcription start site of the Fgf8 gene have been detected

(http://www.ncbi.nlm.nih.gov/epigenomics/). The presence of these CpG islands in the Fgf8 gene suggests that Fgf8 transcription can be influenced by DNA methylation.

Specific aims of dissertation

Based on our data, we conclude that FGF8 signaling is crucial for GnRH neuron development.

However, very little is known about mechanisms driving Fgf8 transcription and the role of important transcription factors highly expressed during GnRH neuron onset. Therefore, the following aims will test the ability of FGF8 to be controlled either epigenetically, or by transcription factors, which ultimately leads to the onset of GnRH neurons.

Aim 1. Test the hypothesis AR and RAR transcription factors drive Fgf8 transcription.

Rationale: It is unclear how Fgf8 transcription is controlled in the mammalian OP. Data from breast and prostatic cancer cell studies indicate that Fgf8 transcription is, in part, under the regulatory control of androgen signaling through androgen receptors (AR) (Evans, 1988;

Gnanapragasam et al., 2002; Ohuchi et al., 1994; Tanaka et al., 2002; Tanaka et al., 1992;

Yamanishi et al., 1995). Specifically, androgen-treatment induced, whereas the androgen antagonist, bicalutamide, inhibited FGF8 protein, and mRNA expression in AR-positive mouse mammary Shionogi carcinoma (SC)-3 cells (Tanaka et al., 1992; Yamanishi et al., 1995).

Similarly, androgen signaling increased Fgf8 mRNA expression in AR positive human prostate

LNCaP cells (Gnanapragasam et al., 2002; Yamanishi et al., 1995). These data indicate that androgen signaling may be a general cellular mechanism that induces Fgf8 expression levels.

Similarly, during early embryonic development, retinoic acid (RA) is secreted from mesodermal tissue and is involved in the induction of neuronal differentiation, apoptosis, proliferation as well as retinal development (Das et al., 2014; Dubey et al., 2018; Mey et al.,

1997; Pourjafar et al., 2017). During patterning of vertebrate trunk development, RA and FGF signaling have opposing diffuse signaling patterns. In the developing, embryonic day (E) 7.5 mouse embryo, RA is secreted in the presomatic mesoderm, and can indirectly influence trunk development by repressing Fgf8 expression during body axis extension (Kumar & Duester,

2014). Duester et. al. showed that the promoter region of FGF8 contains a conserved retinoic acid response element (RARE), which binds retinoic acid receptors (RAR) and recruits repressive histone modifications to transcriptionally repress Fgf8 (Kumar & Duester, 2014). RA has also been shown to restrict Fgf8 expression to the primitive ectoderm, preventing Fgf8 expression in the neuroectoderm (Kumar & Duester, 2014; Sirbu & Duester, 2006). While these

17 experiments are seminal in demonstrating the ability of RA to regulate Fgf8 transcription and ultimately, embryonic patterning, it is currently unknown if RA also regulates Fgf8 in the olfactory placode during GnRH neuron onset.

In these studies, we hypothesized that AR or RA will directly bind to their response elements on the Fgf8 promoter during OP development and modulate the both expression of

Fgf8 mRNA and therefore control the emergence of GnRH neurons from the olfactory placode.

Here, we performed experiments to determine the molecular control of Fgf8 as well as GnRH transcription in the presence or absence of AR or RA. We also analyzed the functional effects of

RA on GnRH neuron cell migration and proliferation using a cell scraping assay.

Aim 2. Test the hypothesis that temporal regulation of Fgf8 transcription is under epigenetic control.

Rationale: Timing of gene expression during embryonic brain development is imperative during fate specification of neuronal populations, of which Fgf8 is critically important. During development of the OP, we found high expression in Tet1-3 mRNA at E13.5. Furthermore, dot blot analysis showed higher levels of 5hmC at E10.5 than E13.5, indicating a genome-wide demethylation event. In neurons, the transition from neural progenitor to differentiation, is associated with high levels of 5hmC (M. A. Hahn et al., 2013; Santiago et al., 2014; Sun et al.,

2014). At the genomic level, it is well known that 5hmC associates with genes important for neuronal function and correspond with gene transcription (M. A. Hahn et al., 2013; Lister et al.,

2013; Szulwach et al., 2011). However, it is unclear how neural OP progenitor cells coordinate the process of DNA demethylation of the Fgf8 locus during development and what factors are required for this process. Aberrant Fgf8 transcription during OP development could result in a

KS like phenotype, where GnRH neurons are reduced or absent. Additionally, Fgf8 transcription is important for formation of forebrain and midbrain-hindbrain structures, which may also be temporally regulated through an analogous epigenetic mechanism. This aim will test the contribution of two epigenetic mechanisms (i.e., TET1 and PRC-2) on timing of transcription of

Fgf8 during emergence of GnRH neurons.

Aim 3. Test the hypothesis that TET1 knockout mice have impaired GnRH neuron development.

Rationale: Our prior studies demonstrated that Tet1 is required for normal Fgf8 transcription in

GnRH neurons, which suggests that TET1 may also play a role during GnRH neuron emergence.

Indeed, global Tet1 and/or Tet2 null mice exhibit disrupted reproduction. Furthermore, Tet1-/- mice exhibit abnormal ovarian follicle development (Dawlaty et al., 2013; Dawlaty et al., 2011).

Unfortunately, no data are available for GnRH neuronal development, function or gonadal steroid hormone levels. Given our preliminary in vitro data, we predict that Tet1 null mice have aberrant GnRH neuron development (i.e., emergence), which may account for fertility deficits.

REFERENCES

Albuisson, J., Pecheux, C., Carel, J. C., Lacombe, D., Leheup, B., Lapuzina, P., . . . Dode, C.

(2005). Kallmann syndrome: 14 novel mutations in KAL1 and FGFR1 (KAL2). Hum

Mutat, 25(1), 98-99.

Alexandre, P., Bachy, I., Marcou, M., & Wassef, M. (2006). Positive and negative regulations by

FGF8 contribute to midbrain roof plate developmental plasticity. Development, 133(15),

2905-2913. doi: 10.1242/dev.02460

Aylwin, C. F., Toro, C. A., Shirtcliff, E., & Lomniczi, A. (2019). Emerging Genetic and

Epigenetic Mechanisms Underlying Pubertal Maturation in Adolescence. J Res Adolesc,

29(1), 54-79. doi: 10.1111/jora.12385

Bachler, M., & Neubuser, A. (2001). Expression of members of the Fgf family and their

receptors during midfacial development. Mech Dev, 100(2), 313-316.

Bernstein, B. E., Mikkelsen, T. S., Xie, X., Kamal, M., Huebert, D. J., Cuff, J., . . . Lander, E. S.

(2006). A bivalent chromatin structure marks key developmental genes in embryonic

stem cells. Cell, 125(2), 315-326. doi: 10.1016/j.cell.2006.02.041

Bird, A. (2007). Perceptions of epigenetics. Nature, 447(7143), 396-398.

Bird, A. P., & Wolffe, A. P. (1999). Methylation-induced repression--belts, braces, and

chromatin. Cell, 99(5), 451-454.

Boehm, U., Bouloux, P. M., Dattani, M. T., de Roux, N., Dode, C., Dunkel, L., . . . Young, J.

(2015). Expert consensus document: European Consensus Statement on congenital

hypogonadotropic hypogonadism--pathogenesis, diagnosis and treatment. Nat Rev

Endocrinol, 11(9), 547-564. doi: 10.1038/nrendo.2015.112

Branco, M. R., Ficz, G., & Reik, W. (2011). Uncovering the role of 5-hydroxymethylcytosine in

the epigenome. Nat Rev Genet, 13(1), 7-13. doi: 10.1038/nrg3080

Brooks, L. R., Chung, W. C., & Tsai, P. S. (2010). Abnormal hypothalamic oxytocin system in

fibroblast growth factor 8-deficient mice. Endocrine, 38(2), 174-180. doi:

10.1007/s12020-010-9366-9

Brooks, L. R., Le, C. D., Chung, W. C., & Tsai, P. S. (2012). Maternal behavior in transgenic

mice with reduced fibroblast growth factor receptor function in gonadotropin-releasing

hormone neurons. Behav Brain Funct, 8, 47. doi: 10.1186/1744-9081-8-47

Cartron, P. F., Nadaradjane, A., Lepape, F., Lalier, L., Gardie, B., & Vallette, F. M. (2013).

Identification of TET1 Partners That Control Its DNA-Demethylating Function. Genes

Cancer, 4(5-6), 235-241. doi: 10.1177/1947601913489020

Chamberlain, S. J., Yee, D., & Magnuson, T. (2008). Polycomb repressive complex 2 is

dispensable for maintenance of embryonic stem cell pluripotency. Stem Cells, 26(6),

1496-1505. doi: 10.1634/stemcells.2008-0102

Cho, H. J., Shan, Y., Whittington, N. C., & Wray, S. (2019). Nasal Placode Development, GnRH

Neuronal Migration and Kallmann Syndrome. Front Cell Dev Biol, 7, 121. doi:

10.3389/fcell.2019.00121

Cho, S., Cho, H., Geum, D., & Kim, K. (1998). Retinoic acid regulates gonadotropin-releasing

hormone (GnRH) release and gene expression in the rat hypothalamic fragments and

GT1-1 neuronal cells in vitro. Brain Res Mol Brain Res, 54(1), 74-84. doi:

10.1016/s0169-328x(97)00325-2

Chung, W. C., & Auger, A. P. (2013). Gender differences in neurodevelopment and epigenetics.

Pflugers Arch, 465(5), 573-584. doi: 10.1007/s00424-013-1258-4

Chung, W. C., Linscott, M. L., Rodriguez, K. M., & Stewart, C. E. (2016). The Regulation and

Function of Fibroblast Growth Factor 8 and Its Function during Gonadotropin-Releasing

Hormone Neuron Development. Front Endocrinol (Lausanne), 7, 114. doi:

10.3389/fendo.2016.00114

Chung, W. C., Matthews, T. A., Tata, B. K., & Tsai, P. S. (2010). Compound deficiencies in

multiple fibroblast growth factor signalling components differentially impact the murine

gonadotrophin-releasing hormone system. J Neuroendocrinol, 22(8), 944-950. doi:

10.1111/j.1365-2826.2010.02024.x

Chung, W. C., Moyle, S. S., & Tsai, P. S. (2008). Fibroblast growth factor 8 signaling through

fibroblast growth factor receptor 1 is required for the emergence of gonadotropin-

releasing hormone neurons. Endocrinology, 149(10), 4997-5003. doi: 10.1210/en.2007-

1634

Chung, W. C., & Tsai, P. S. (2010). Role of fibroblast growth factor signaling in gonadotropin-

releasing hormone neuronal system development. Front Horm Res, 39, 37-50. doi:

10.1159/000312692

Cole LW, H. F., Seminara SS, Hughes VA, Dwyer A, Quinton R, Pearce SH, Hall JE, Crowley

WF,Mohammadi M, Pitteloud N. (2007). FGF8 Is a Key Ligand for FGFR1 in GnRH

Ontogeny: Evidence from a Human Disease Model.

Cole, L. W., Sidis, Y., Zhang, C., Quinton, R., Plummer, L., Pignatelli, D., . . . Pitteloud, N.

(2008). Mutations in prokineticin 2 and prokineticin receptor 2 genes in human

gonadotrophin-releasing hormone deficiency: molecular genetics and clinical spectrum. J

Clin Endocrinol Metab, 93(9), 3551-3559. doi: 10.1210/jc.2007-2654

Conrad, B., Kriebel, J., & Hetzel, W. (1978). Hereditary bimanual synkinesis combined with

hypogonadotropic hypogonadism and anosmia in four brothers. Journal of neurology,

218(4), 263-274.

Creuzet, S., Schuler, B., Couly, G., & Le Douarin, N. M. (2004). Reciprocal relationships

between Fgf8 and neural crest cells in facial and forebrain development. Proc Natl Acad

Sci U S A, 101(14), 4843-4847. doi: 10.1073/pnas.0400869101

Cunningham, T. J., Brade, T., Sandell, L. L., Lewandoski, M., Trainor, P. A., Colas, A., . . .

Duester, G. (2015). Retinoic Acid Activity in Undifferentiated Neural Progenitors Is

Sufficient to Fulfill Its Role in Restricting Fgf8 Expression for Somitogenesis. PLoS One,

10(9), e0137894. doi: 10.1371/journal.pone.0137894

Das, B. C., Thapa, P., Karki, R., Das, S., Mahapatra, S., Liu, T. C., . . . Evans, T. (2014).

Retinoic acid signaling pathways in development and diseases. Bioorg Med Chem, 22(2),

673-683. doi: 10.1016/j.bmc.2013.11.025

Dawlaty, M. M., Breiling, A., Le, T., Raddatz, G., Barrasa, M. I., Cheng, A. W., . . . Jaenisch, R.

(2013). Combined deficiency of Tet1 and Tet2 causes epigenetic abnormalities but is

compatible with postnatal development. Dev Cell, 24(3), 310-323. doi:

10.1016/j.devcel.2012.12.015

Dawlaty, M. M., Ganz, K., Powell, B. E., Hu, Y. C., Markoulaki, S., Cheng, A. W., . . . Jaenisch,

R. (2011). Tet1 is dispensable for maintaining pluripotency and its loss is compatible

with embryonic and postnatal development. Cell Stem Cell, 9(2), 166-175. doi:

10.1016/j.stem.2011.07.010

Denis, H., Ndlovu, M. N., & Fuks, F. (2011). Regulation of mammalian DNA

methyltransferases: a route to new mechanisms. EMBO Reports, 12(7), 647-656. doi:

10.1038/embor.2011.110

Dode, C., Fouveaut, C., Mortier, G., Janssens, S., Bertherat, J., Mahoudeau, J., . . . Hardelin, J. P.

(2007). Novel FGFR1 sequence variants in Kallmann syndrome, and genetic evidence

that the FGFR1c isoform is required in olfactory bulb and palate morphogenesis. Hum

Mutat, 28(1), 97-98. doi: 10.1002/humu.9470

Dode, C., Levilliers, J., Dupont, J. M., De Paepe, A., Le Du, N., Soussi-Yanicostas, N., . . .

Hardelin, J. P. (2003). Loss-of-function mutations in FGFR1 cause autosomal dominant

Kallmann syndrome. Nat Genet, 33(4), 463-465. doi: 10.1038/ng1122

Dode, C., Teixeira, L., Levilliers, J., Fouveaut, C., Bouchard, P., Kottler, M. L., . . . Hardelin, J.

P. (2006). Kallmann syndrome: mutations in the genes encoding prokineticin-2 and

prokineticin receptor-2. PLoS Genet, 2(10), e175.

Dubey, A., Rose, R. E., Jones, D. R., & Saint-Jeannet, J. P. (2018). Generating retinoic acid

gradients by local degradation during craniofacial development: One cell's cue is another

cell's poison. Genesis, 56(2). doi: 10.1002/dvg.23091

Evans, R. M. (1988). The steroid and thyroid hormone receptor superfamily. Science, 240(4854),

889-895. doi: 10.1126/science.3283939

Falardeau, J., Chung, W. C., Beenken, A., Raivio, T., Plummer, L., Sidis, Y., . . . Pitteloud, N.

(2008). Decreased FGF8 signaling causes deficiency of gonadotropin-releasing hormone

in humans and mice. J Clin Invest, 118(8), 2822-2831. doi: 10.1172/jci34538

Fisher, C. L., & Fisher, A. G. (2011). Chromatin states in pluripotent, differentiated, and

reprogrammed cells. Curr Opin Genet Dev, 21(2), 140-146. doi:

10.1016/j.gde.2011.01.015

Forni, P. E., Bharti, K., Flannery, E. M., Shimogori, T., & Wray, S. (2013). The indirect role of

fibroblast growth factor-8 in defining neurogenic niches of the olfactory/GnRH systems.

J Neurosci, 33(50), 19620-19634. doi: 10.1523/jneurosci.3238-13.2013

Forni, P. E., Taylor-Burds, C., Melvin, V. S., Williams, T., & Wray, S. (2011). Neural crest and

ectodermal cells intermix in the nasal placode to give rise to GnRH-1 neurons, sensory

neurons, and olfactory ensheathing cells. J Neurosci, 31(18), 6915-6927. doi:

10.1523/jneurosci.6087-10.2011

Forni, P. E., & Wray, S. (2015). GnRH, anosmia and hypogonadotropic hypogonadism--where

are we? Front Neuroendocrinol, 36, 165-177. doi: 10.1016/j.yfrne.2014.09.004

Franco, B., Guioli, S., Pragliola, A., Incerti, B., Bardoni, B., Tonlorenzi, R., . . . Ballabio, A.

(1991). A gene deleted in Kallmann's syndrome shares homology with neural cell

adhesion and axonal path-finding molecules. Nature, 353(6344), 529-536. doi:

10.1038/353529a0

Gill, J. C., Moenter, S. M., & Tsai, P. S. (2004). Developmental regulation of gonadotropin-

releasing hormone neurons by fibroblast growth factor signaling. Endocrinology, 145(8),

3830-3839. doi: 10.1210/en.2004-0214

Gill, J. C., & Tsai, P. S. (2006). Expression of a dominant negative FGF receptor in developing

GNRH1 neurons disrupts axon outgrowth and targeting to the median eminence. Biol

Reprod, 74(3), 463-472.

Gnanapragasam, V. J., Robson, C. N., Neal, D. E., & Leung, H. Y. (2002). Regulation of FGF8

expression by the androgen receptor in human prostate cancer. Oncogene, 21(33), 5069-

5080. doi: 10.1038/sj.onc.1205663

Grafstrom, R. H., Yuan, R., & Hamilton, D. L. (1985). The characteristics of DNA methylation

in an in vitro DNA synthesizing system from mouse fibroblasts. Nucleic Acids Res.,

13(8), 2827-2842.

Grandy, R. A., Whitfield, T. W., Wu, H., Fitzgerald, M. P., VanOudenhove, J. J., Zaidi, S. K., . .

. Stein, G. S. (2015). Genome-Wide Studies Reveal that H3K4me3 Modification in

Bivalent Genes Is Dynamically Regulated during the Pluripotent Cell Cycle and

Stabilized upon Differentiation. Mol Cell Biol, 36(4), 615-627. doi: 10.1128/mcb.00877-

Gray, J. D., Kogan, J. F., Marrocco, J., & McEwen, B. S. (2017). Genomic and epigenomic

mechanisms of glucocorticoids in the brain. Nat Rev Endocrinol, 13(11), 661-673. doi:

10.1038/nrendo.2017.97

Hahn, Maria A., Qiu, R., Wu, X., Li, Arthur X., Zhang, H., Wang, J., . . . Lu, Q. (2013).

Dynamics of 5-Hydroxymethylcytosine and Chromatin Marks in Mammalian

Neurogenesis. Cell Rep, 3(2), 291-300. doi: https://doi.org/10.1016/j.celrep.2013.01.011

Hahn, M. A., Qiu, R., Wu, X., Li, A. X., Zhang, H., Wang, J., . . . Lu, Q. (2013). Dynamics of 5-

hydroxymethylcytosine and chromatin marks in Mammalian neurogenesis. Cell Rep,

3(2), 291-300. doi: 10.1016/j.celrep.2013.01.011

Hardelin, J.-P., Levilliers, J., Blanchard, S., Carel, J.-C., Leutenegger, M., Pinard-Bertelletto, J.-

P., . . . Petit, C. (1993). Heterogeneity in the mutations responsible for X chromosome-

linked Kallmann syndrome. Hum Mol Genet, 2(4), 373-377.

Hardelin, J., Levilliers, J., Young, J., Pholsena, M., Legouis, R., Kirk, J., . . . Schaison, G.

(1993). Xp22. 3 deletions in isolated familial Kallmann's syndrome. The Journal of

Clinical Endocrinology & Metabolism, 76(4), 827-831.

Hardelin, J. P., Levilliers, J., del Castillo, I., Cohen-Salmon, M., Legouis, R., Blanchard, S., . . .

et al. (1992). X chromosome-linked Kallmann syndrome: stop mutations validate the

candidate gene. Proc Natl Acad Sci U S A, 89(17), 8190-8194.

Hirabayashi, Y., Suzki, N., Tsuboi, M., Endo, T. A., Toyoda, T., Shinga, J., . . . Gotoh, Y.

(2009). Polycomb limits the neurogenic competence of neural precursor cells to promote

astrogenic fate transition. Neuron, 63(5), 600-613. doi: 10.1016/j.neuron.2009.08.021

Hoch, R. V., Clarke, J. A., & Rubenstein, J. L. (2015). Fgf signaling controls the telencephalic

distribution of Fgf-expressing progenitors generated in the rostral patterning center.

Neural Dev, 10, 8. doi: 10.1186/s13064-015-0037-7

Huang, Y., Chavez, L., Chang, X., Wang, X., Pastor, W. A., Kang, J., . . . Rao, A. (2014).

Distinct roles of the methylcytosine oxidases Tet1 and Tet2 in mouse embryonic stem

cells. Proc Natl Acad Sci U S A, 111(4), 1361-1366. doi: 10.1073/pnas.1322921111

Ito, S., D'Alessio, A. C., Taranova, O. V., Hong, K., Sowers, L. C., & Zhang, Y. (2010). Role of

Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass

specification. Nature, 466(7310), 1129-1133. doi: 10.1038/nature09303

Iyer, A. K., Brayman, M. J., & Mellon, P. L. (2011). Dynamic chromatin modifications control

GnRH gene expression during neuronal differentiation and protein kinase C signal

transduction. Molecular endocrinology (Baltimore, Md.), 25(3), 460-473. doi:

10.1210/me.2010-0403

Katoh, H., Shibata, S., Fukuda, K., Sato, M., Satoh, E., Nagoshi, N., . . . Okano, H. (2011). The

dual origin of the peripheral olfactory system: placode and neural crest. Mol Brain, 4, 34.

doi: 10.1186/1756-6606-4-34

Kawauchi, S., Shou, J., Santos, R., Hebert, J. M., McConnell, S. K., Mason, I., & Calof, A. L.

(2005). Fgf8 expression defines a morphogenetic center required for olfactory

neurogenesis and nasal cavity development in the mouse. Development, 132(23), 5211-

5223. doi: 10.1242/dev.02143

Kawauchi, S., Young, C., & Calof, A. (2007). Requirement for Fgf8 in generation of GnRH-

expressing hypothalamic neurons during mouse development. Annual Meeting, Society

for Neuroscience, 33.

Kim, H. G., Kurth, I., Lan, F., Meliciani, I., Wenzel, W., Eom, S. H., . . . Layman, L. C. (2008).

Mutations in CHD7, encoding a chromatin-remodeling protein, cause idiopathic

hypogonadotropic hypogonadism and Kallmann syndrome. Am J Hum Genet, 83(4), 511-

519.

Klose, R. J., & Bird, A. P. (2006). Genomic DNA methylation: the mark and its mediators.

Trends Biochem.Sci., 31(2), 89-97.

Kong, L., Tan, L., Lv, R., Shi, Z., Xiong, L., Wu, F., . . . Shi, Y. G. (2016). A primary role of

TET proteins in establishment and maintenance of De Novo bivalency at CpG islands.

Nucleic Acids Res, 44(18), 8682-8692. doi: 10.1093/nar/gkw529

Kouhara, H., Koga, M., Kasayama, S., Tanaka, A., Kishimoto, T., & Sato, B. (1994).

Transforming activity of a newly cloned androgen-induced growth factor. Oncogene,

9(2), 455-462.

Kramer, P. R., & Wray, S. (2000). Midline nasal tissue influences nestin expression in nasal-

placode-derived luteinizing hormone-releasing hormone neurons during development.

Dev Biol, 227(2), 343-357. doi: 10.1006/dbio.2000.9896

Kramer, P. R., & Wray, S. (2001). Nasal embryonic LHRH factor (NELF) expression within the

CNS and PNS of the rodent. Brain Res Gene Expr Patterns, 1(1), 23-26.

Kumar, S., & Duester, G. (2014). Retinoic acid controls body axis extension by directly

repressing Fgf8 transcription. Development, 141(15), 2972-2977. doi:

10.1242/dev.112367

Kurian, J. R., Louis, S., Keen, K. L., Wolfe, A., Terasawa, E., & Levine, J. E. (2016). The

Methylcytosine Dioxygenase Ten-Eleven Translocase-2 (tet2) Enables Elevated GnRH

Gene Expression and Maintenance of Male Reproductive Function. Endocrinology,

157(9), 3588-3603. doi: 10.1210/en.2016-1087

Kurian, J. R., & Terasawa, E. (2013). Epigenetic control of gonadotropin releasing hormone

neurons. Front Endocrinol (Lausanne), 4, 61. doi: 10.3389/fendo.2013.00061

Laitinen, E. M., Vaaralahti, K., Tommiska, J., Eklund, E., Tervaniemi, M., Valanne, L., &

Raivio, T. (2011). Incidence, phenotypic features and molecular genetics of Kallmann

syndrome in Finland. Orphanet J Rare Dis, 6, 41. doi: 10.1186/1750-1172-6-41

Laverriere, J. N., L'Hote, D., Tabouy, L., Schang, A. L., Querat, B., & Cohen-Tannoudji, J.

(2016). Epigenetic regulation of alternative promoters and enhancers in progenitor,

immature, and mature gonadotrope cell lines. Mol Cell Endocrinol, 434, 250-265. doi:

10.1016/j.mce.2016.07.010

Legouis, R., Hardelin, J. P., Levilliers, J., Claverie, J. M., Compain, S., Wunderle, V., . . . et al.

(1991). The candidate gene for the X-linked Kallmann syndrome encodes a protein

related to adhesion molecules. Cell, 67(2), 423-435.

Leung, H. Y., Dickson, C., Robson, C. N., & Neal, D. E. (1996). Over-expression of fibroblast

growth factor-8 in human prostate cancer. Oncogene, 12(8), 1833-1835.

Linscott, M. L., & Chung, W. C. J. (2019). TET1 regulates fibroblast growth factor 8

transcription in gonadotropin releasing hormone neurons. PLoS One, 14(7), e0220530.

doi: 10.1371/journal.pone.0220530

Lister, R., Mukamel, E. A., Nery, J. R., Urich, M., Puddifoot, C. A., Johnson, N. D., . . . Ecker, J.

R. (2013). Global Epigenomic Reconfiguration During Mammalian Brain Development.

Science, 341(6146). doi: 10.1126/science.1237905

Lomniczi, A., Loche, A., Castellano, J. M., Ronnekleiv, O. K., Bosch, M., Kaidar, G., . . . Ojeda,

S. R. (2013). Epigenetic control of female puberty. Nat Neurosci, 16(3), 281-289. doi:

10.1038/nn.3319

Lomniczi, A., & Ojeda, S. R. (2016). The Emerging Role of Epigenetics in the Regulation of

Female Puberty. Endocr Dev, 29, 1-16. doi: 10.1159/000438840

Luca, T., & Comoglio Paolo, M. (2004). To move or not to move? Semaphorin signalling in cell

migration. EMBO Rep, 5(4), 356-362.

Marinic, M., Aktas, T., Ruf, S., & Spitz, F. (2013). An integrated holo-enhancer unit defines

tissue and gene specificity of the Fgf8 regulatory landscape. Dev Cell, 24(5), 530-542.

doi: 10.1016/j.devcel.2013.01.025

Mascarenhas, M. N., Flaxman, S. R., Boerma, T., Vanderpoel, S., & Stevens, G. A. (2012).

National, regional, and global trends in infertility prevalence since 1990: a systematic

analysis of 277 health surveys. PLoS Med, 9(12), e1001356. doi:

10.1371/journal.pmed.1001356

Mey, J., McCaffery, P., & Dräger, U. C. (1997). Retinoic Acid Synthesis in the Developing

Chick Retina. The Journal of Neuroscience, 17(19), 7441-7449. doi:

10.1523/jneurosci.17-19-07441.1997

Miraoui, H., Dwyer, A. A., Sykiotis, G. P., Plummer, L., Chung, W., Feng, B., . . . Pitteloud, N.

(2013). Mutations in FGF17, IL17RD, DUSP6, SPRY4, and FLRT3 are identified in

individuals with congenital hypogonadotropic hypogonadism. Am J Hum Genet, 92(5),

725-743. doi: 10.1016/j.ajhg.2013.04.008

Mølsted, K., Kjær, I., Giwercman, A., Vesterhauge, S., & Skakkebæk, N. E. (1997). Craniofacial

morphology in patients with Kallmann's syndrome with and without cleft lip and palate.

The Cleft palate-craniofacial journal, 34(5), 417-424.

Münzel, M., Globisch, D., Brückl, T., Wagner, M., Welzmiller, V., Michalakis, S., . . . Carell, T.

(2010). Quantification of the Sixth DNA Base Hydroxymethylcytosine in the Brain.

Angewandte Chemie International Edition, 49(31), 5375-5377. doi:

10.1002/anie.201002033

Murakami, S., Kikuyama, S., & Arai, Y. (1992). The origin of the luteinizing hormone-releasing

hormone (LHRH) neurons in newts (Cynops pyrrhogaster): the effect of olfactory

placode ablation. Cell Tissue Res, 269(1), 21-27.

Neri, F., Incarnato, D., Krepelova, A., Rapelli, S., Pagnani, A., Zecchina, R., . . . Oliviero, S.

(2013). Genome-wide analysis identifies a functional association of Tet1 and Polycomb

repressive complex 2 in mouse embryonic stem cells. Genome Biol, 14(8), R91. doi:

10.1186/gb-2013-14-8-r91

Northcutt, R. G., & Muske, L. E. (1994). Multiple embryonic origins of gonadotropin-releasing

hormone (GnRH) immunoreactive neurons. Brain Res Dev Brain Res, 78(2), 279-290.

doi: 10.1016/0165-3806(94)90037-x

Ohuchi, H., Yoshioka, H., Tanaka, A., Kawakami, Y., Nohno, T., & Noji, S. (1994).

Involvement of androgen-induced growth factor (FGF-8) gene in mouse embryogenesis

and morphogenesis. Biochem Biophys Res Commun, 204(2), 882-888. doi:

10.1006/bbrc.1994.2542

Ornitz, D. M., & Itoh, N. (2015). The Fibroblast Growth Factor signaling pathway. Wiley

Interdiscip Rev Dev Biol, 4(3), 215-266. doi: 10.1002/wdev.176

Ornitz, D. M., Xu, J., Colvin, J. S., McEwen, D. G., MacArthur, C. A., Coulier, F., . . . Goldfarb,

M. (1996). Receptor Specificity of the Fibroblast Growth Factor Family. J. Biol. Chem.,

271(25), 15292-15297. doi: 10.1074/jbc.271.25.15292

Pasini, D., Bracken, A. P., Hansen, J. B., Capillo, M., & Helin, K. (2007). The polycomb group

protein Suz12 is required for embryonic stem cell differentiation. Mol Cell Biol, 27(10),

3769-3779. doi: 10.1128/mcb.01432-06

Pereira, J. D., Sansom, S. N., Smith, J., Dobenecker, M. W., Tarakhovsky, A., & Livesey, F. J.

(2010). Ezh2, the histone methyltransferase of PRC2, regulates the balance between self-

renewal and differentiation in the cerebral cortex. Proc Natl Acad Sci U S A, 107(36),

15957-15962. doi: 10.1073/pnas.1002530107

Pitteloud, N. (2007). New Insights into the Genetics of Kallmann Syndrome. Annual Meeting,

Endocrine Society, 90.

Pitteloud, N., Quinton, R., Pearce, S., Raivio, T., Acierno, J., Dwyer, A., . . . Crowley, W.

(2007). Digenic mutations account for variable phenotypes in idiopathic

hypogonadotropic hypogonadism. J Clin Invest, 117(2), 457-463.

Pourjafar, M., Saidijam, M., Etemadi, K., & Najafi, R. (2017). All-trans retinoic acid enhances in

vitro mesenchymal stem cells migration by targeting matrix metalloproteinases 2 and 9.

Biotechnol Lett, 39(8), 1263-1268. doi: 10.1007/s10529-017-2350-1

Quinton, R., Duke, V., De Zoysa, P., Platts, A., Valentine, A., Kendall, B., . . . Jacobs, H. (1996).

The neuroradiology of Kallmann's syndrome: a genotypic and phenotypic analysis. The

Journal of Clinical Endocrinology & Metabolism, 81(8), 3010-3017.

Quinton, R., Duke, V. M., Robertson, A., Kirk, J. M., Matfin, G., De Zoysa, P. A., . . . Conway,

G. S. (2001). Idiopathic gonadotrophin deficiency: genetic questions addressed through

phenotypic characterization. Clinical endocrinology, 55(2), 163-174.

Ramsahoye, B. H., Biniszkiewicz, D., Lyko, F., Clark, V., Bird, A. P., & Jaenisch, R. (2000).

Non-CpG methylation is prevalent in embryonic stem cells and may be mediated by

DNA methyltransferase 3a. Proc.Natl.Acad.Sci.U.S.A, 97(10), 5237-5242.

Reardon, W. (2007). Kallmann syndrome presenting as congenital ptosis in brothers. Clinical

dysmorphology, 16(3), 207-208.

Report on optimal evaluation of the infertile male. (2006). Fertil Steril, 86(5 Suppl 1), S202-209.

doi: 10.1016/j.fertnstert.2006.08.029

Rochester, J. R., Chung, W. C., Hayes, T. B., & Tsai, P. S. (2012). Opposite-sex housing

reactivates the declining GnRH system in aged transgenic male mice with FGF signaling

deficiency. Am J Physiol Endocrinol Metab, 303(12), E1428-1439. doi:

10.1152/ajpendo.00289.2012

Rodriguez, K. M., Stevenson, E. L., Stewart, C. E., Linscott, M. L., & Chung, W. C. (2015).

Fibroblast growth factor 8 regulates postnatal development of paraventricular nucleus

neuroendocrine cells. Behav Brain Funct, 11(1), 34. doi: 10.1186/s12993-015-0081-9

Sabado, V., Barraud, P., Baker, C. V., & Streit, A. (2012). Specification of GnRH-1 neurons by

antagonistic FGF and retinoic acid signaling. Dev Biol, 362(2), 254-262. doi:

10.1016/j.ydbio.2011.12.016

Saito, H., Kasayama, S., Kouhara, H., Matsumoto, K., & Sato, B. (1991). Up-regulation of

fibroblast growth factor (FGF) receptor mRNA levels by basic FGF or testosterone in

androgen-sensitive mouse mammary tumor cells. Biochem Biophys Res Commun, 174(1),

136-141.

Santiago, M., Antunes, C., Guedes, M., Sousa, N., & Marques, C. J. (2014). TET enzymes and

DNA hydroxymethylation in neural development and function - how critical are they?

Genomics, 104(5), 334-340. doi: 10.1016/j.ygeno.2014.08.018

Sarfati, J., Guiochon-Mantel, A., Rondard, P., Arnulf, I., Garcia-Piñero, A., Wolczynski, S., . . .

Mathieu, M. (2010). A comparative phenotypic study of kallmann syndrome patients

carrying monoallelic and biallelic mutations in the prokineticin 2 or prokineticin receptor

2 genes. The Journal of Clinical Endocrinology & Metabolism, 95(2), 659-669.

Schwanzel-Fukuda, M., Bick, D., & Pfaff, D. W. (1989). Luteinizing hormone-releasing

hormone (LHRH)-expressing cells do not migrate normally in an inherited hypogonadal

(Kallmann) syndrome. Mol Brain Res, 6(4), 311-326.

Schwanzel-Fukuda, M., & Pfaff, D. W. (1989). Origin of luteinizing hormone-releasing hormone

neurons. Nature, 338(6211), 161-164. doi: 10.1038/338161a0

Shen, Y., Yue, F., McCleary, D. F., Ye, Z., Edsall, L., Kuan, S., . . . Ren, B. (2012). A map of

the cis-regulatory sequences in the mouse genome. Nature, 488(7409), 116-120. doi:

10.1038/nature11243

Sidhoum, V. F., Chan, Y. M., Lippincott, M. F., Balasubramanian, R., Quinton, R., Plummer, L.,

. . . Seminara, S. B. (2014). Reversal and relapse of hypogonadotropic hypogonadism:

resilience and fragility of the reproductive neuroendocrine system. J Clin Endocrinol

Metab, 99(3), 861-870. doi: 10.1210/jc.2013-2809

Sirbu, I. O., & Duester, G. (2006). Retinoic-acid signalling in node ectoderm and posterior

neural plate directs left-right patterning of somitic mesoderm. Nat Cell Biol, 8(3), 271-

277. doi: 10.1038/ncb1374

Su, P. Y., Xu, G., Han, A. Z., Xu, N., Zhang, G. B., & Tao, F. B. (2019). [The role of

hypothalamus polycomb gene methylation in bisphenol A exposure during pregnancy and

premature puberty in female offspring]. Zhonghua Yu Fang Yi Xue Za Zhi, 53(3), 303-

308. doi: 10.3760/cma.j.issn.0253-9624.2019.03.013

Sun, W., Zang, L., Shu, Q., & Li, X. (2014). From development to diseases: the role of 5hmC in

brain. Genomics, 104(5), 347-351. doi: 10.1016/j.ygeno.2014.08.021

Szulwach, K. E., Li, X., Li, Y., Song, C. X., Wu, H., Dai, Q., . . . Jin, P. (2011). 5-hmC-mediated

epigenetic dynamics during postnatal neurodevelopment and aging. Nat Neurosci, 14(12),

1607-1616. doi: 10.1038/nn.2959

Tanaka, A., Kamiakito, T., Takayashiki, N., Sakurai, S., & Saito, K. (2002). Fibroblast growth

factor 8 expression in breast carcinoma: associations with androgen receptor and

prostate-specific antigen expressions. Virchows Arch, 441(4), 380-384. doi:

10.1007/s00428-002-0687-9

Tanaka, A., Miyamoto, K., Minamino, N., Takeda, M., Sato, B., Matsuo, H., & Matsumoto, K.

(1992). Cloning and characterization of an androgen-induced growth factor essential for

the androgen-dependent growth of mouse mammary carcinoma cells. Proc Natl Acad Sci

U S A, 89(19), 8928-8932. doi: 10.1073/pnas.89.19.8928

Tata, B. K., Chung, W. C., Brooks, L. R., Kavanaugh, S. I., & Tsai, P. S. (2012). Fibroblast

growth factor signaling deficiencies impact female reproduction and kisspeptin neurons

in mice. Biol Reprod, 86(4), 119. doi: 10.1095/biolreprod.111.095992

Tsai, P. S., Brooks, L. R., Rochester, J. R., Kavanaugh, S. I., & Chung, W. C. (2011). Fibroblast

growth factor signaling in the developing neuroendocrine hypothalamus. Front

Neuroendocrinol, 32(1), 95-107. doi: 10.1016/j.yfrne.2010.11.002

Tsai, P. S., & Gill, J. C. (2006). Mechanisms of disease: Insights into X-linked and autosomal-

dominant Kallmann syndrome. Nat Clin Pract Endocrinol Metab, 2(3), 160-171. doi:

10.1038/ncpendmet0119

Tsai, P. S., Moenter, S. M., Postigo, H. R., El Majdoubi, M., Pak, T. R., Gill, J. C., . . . Weiner,

R. I. (2005). Targeted expression of a dominant-negative fibroblast growth factor (FGF)

receptor in gonadotropin-releasing hormone (GnRH) neurons reduces FGF

responsiveness and the size of GnRH neuronal population. Mol Endocrinol, 19(1), 225-

236. doi: 10.1210/me.2004-0330

Verma, N., Pan, H., Dore, L. C., Shukla, A., Li, Q. V., Pelham-Webb, B., . . . Huangfu, D.

(2018). TET proteins safeguard bivalent promoters from de novo methylation in human

embryonic stem cells. Nat Genet, 50(1), 83-95. doi: 10.1038/s41588-017-0002-y

Voigt, P., Tee, W. W., & Reinberg, D. (2013). A double take on bivalent promoters. Genes Dev,

27(12), 1318-1338. doi: 10.1101/gad.219626.113

Wegenke, J., Uehling, D., Wear Jr, J., Gordon, E., Bargman, J., Deacon, J., . . . Opitz, J. (1975).

Familial Kallmann syndrome with unilateral renal aplasia. Clinical genetics, 7(5), 368-

381.

Whitlock, K. E., & Westerfield, M. (2000). The olfactory placodes of the zebrafish form by

convergence of cellular fields at the edge of the neural plate. Development, 127(17),

3645-3653.

Wray, S., Grant, P., & Gainer, H. (1989). Evidence that cells expressing luteinizing hormone-

releasing hormone mRNA in the mouse are derived from progenitor cells in the olfactory

placode. Proc Natl Acad Sci U S A, 86(20), 8132-8136.

Wray, S., Nieburgs, A., & Elkabes, S. (1989). Spatiotemporal cell expression of luteinizing

hormone-releasing hormone in the prenatal mouse: evidence for an embryonic origin in

the olfactory placode. Brain Res Dev Brain Res, 46(2), 309-318.

Wu, H., D'Alessio, A. C., Ito, S., Xia, K., Wang, Z., Cui, K., . . . Zhang, Y. (2011). Dual

functions of Tet1 in transcriptional regulation in mouse embryonic stem cells. Nature,

473(7347), 389-393. doi: 10.1038/nature09934

Wu, H., & Zhang, Y. (2011). Mechanisms and functions of Tet protein-mediated 5-

methylcytosine oxidation. Genes Dev, 25(23), 2436-2452. doi: 10.1101/gad.179184.111

Wu, T. J., Gibson, M. J., Rogers, M. C., & Silverman, A. J. (1997). New observations on the

development of the gonadotropin-releasing hormone system in the mouse. J Neurobiol,

33(7), 983-998. doi: 10.1002/(sici)1097-4695(199712)33:7<983::aid-neu9>3.0.co;2-4

Xu, C., Messina, A., Somm, E., Miraoui, H., Kinnunen, T., Acierno, J., Jr., . . . Pitteloud, N.

(2017). KLB, encoding beta-Klotho, is mutated in patients with congenital

hypogonadotropic hypogonadism. EMBO Mol Med, 9(10), 1379-1397. doi:

10.15252/emmm.201607376

Yamanishi, H., Tanaka, A., Miyamoto, K., Takeda, M., Nishizawa, Y., Koga, M., . . .

Matsumoto, K. (1995). An essential role of androgen-induced growth factor in

glucocorticoid-dependent autocrine loop in Shionogi carcinoma 115 cells. J Steroid

Biochem Mol Biol, 52(1), 49-53. doi: 10.1016/0960-0760(94)00148-f

Yoon, H. G., Chan, D. W., Reynolds, A. B., Qin, J., & Wong, J. (2003). N-CoR mediates DNA

methylation-dependent repression through a methyl CpG binding protein Kaiso.

Mol.Cell, 12(3), 723-734.

Yosefzon, Y., David, C., Tsukerman, A., Pnueli, L., Qiao, S., Boehm, U., & Melamed, P. (2017).

An epigenetic switch repressing Tet1 in gonadotropes activates the reproductive axis.

Proc Natl Acad Sci U S A, 114(38), 10131-10136. doi: 10.1073/pnas.1704393114

Young, J., Xu, C., Papadakis, G. E., Acierno, J. S., Maione, L., Hietamaki, J., . . . Pitteloud, N.

(2019). Clinical Management of Congenital Hypogonadotropic Hypogonadism. Endocr

Rev, 40(2), 669-710. doi: 10.1210/er.2018-00116

Zhang, X., Ibrahimi, O. A., Olsen, S. K., Umemori, H., Mohammadi, M., & Ornitz, D. M.

(2006). Receptor specificity of the fibroblast growth factor family. The complete

mammalian FGF family. J Biol Chem, 281(23), 15694-15700.

CHAPTER II

FIBROBLAST GROWTH FACTOR 8 EXPRESSION IN GT1-7 GNRH-SECRETING NEURONS IS ANDROGEN-INDEPENDENT, BUT CAN BE UPREGULATED BY THE INHIBITION OF DNA METHYLTRANSFERASES This work has been published previously as “ Fibroblast Growth Factor 8 Expression in GT1-7 GnRH-Secreting Neurons Is Androgen-Independent, but Can Be Upregulated by the Inhibition of DNA Methyltransferases.” Linscott et. al, Frontiers in cell and developmental biology, 4, 34. doi:10.3389/fcell.2016.00034

INTRODUCTION

Traditionally, fibroblast growth factor 8 (FGF8) has been studied in the context of a potent morphogen that is required for establishing morphogenetic centers in the developing mammalian neural tube (Crossley & Martin, 1995; Meyers et al., 1998; Sun et al., 1999). Indeed, Fgf8 mRNA and protein is highly expressed in the early embryonic midbrain-hindbrain border, and the anterior neural ridge (Crossley & Martin, 1995;

Kawauchi et al., 2004). Recently, we showed that FGF8 function is also important for the development of gonadotropin-releasing hormone (GnRH)-secreting neurons, which reside within the preoptic-hypothalamus region. Specifically, while GnRH neurons were present in newborn wildtype Fgf8 hypomorphic mice, they were absent in homozygous litter mates (W. C. Chung et al., 2008). Additional studies showed that GnRH neurons were already missing in the E11.5 Fgf8 hypomorphic olfactory placode (OP) (W. C.

Chung et al., 2008; Wray, Grant, et al., 1989) from which the majority of GnRH neurons

39 emerge (Schwanzel-Fukuda & Pfaff, 1989; Wray, Grant, et al., 1989). Together these data clearly support the supposition that FGF8 function is required for vertebrate GnRH neuron development (W. C. Chung et al., 2008; P. S. Tsai et al., 2011).

In vitro explant studies in chicken OP explants further pinpointed that FGF8 function is required for the emergence of GnRH neurons. Normally, GnRH precursor cells in the chicken

OP are specified around the Hamburger and Hamilton (HH) stage 16/17. Treatment of HH15 OP with FGF8 advanced the emergence of GnRH neurons by approximately 24 hours (Sabado et al.,

2012), which was abrogated by a FGF antagonist in HH17 OP (Sabado et al., 2012). These studies not only support the general supposition that FGF8 function is required for fate- specifying GnRH precursor cells, but also that FGF8 only has a narrow window of opportunity to induce the emergence of GnRH neurons.

Although much is already known about the effects of FGF8 function on GnRH neuron development, it is unclear how Fgf8 transcription is controlled in the mammalian OP. Data from breast and prostatic cancer cell studies indicate that Fgf8 transcription is, in part, under the regulatory control of androgen signaling through androgen receptors (AR) (Evans, 1988;

Gnanapragasam et al., 2002; Ohuchi et al., 1994; Tanaka et al., 2002; Yamanishi et al., 1995).

Specifically, androgen-treatment induced, whereas the androgen antagonist, bicalutamide, inhibited FGF8 protein and mRNA expression in AR-positive mouse mammary Shionogi carcinoma (SC)-3 cells (Tanaka et al., 1992; Yamanishi et al., 1995). Similarly, androgen signaling increased Fgf8 mRNA expression in AR positive human prostate LNCaP cells

(Gnanapragasam et al., 2002; Yamanishi et al., 1995). These data indicate that androgen signaling may be a general cellular mechanism that induces Fgf8 expression levels.

In these studies, we tested whether androgen is able to induce Fgf8 transcription in GnRH neurons. For this purpose, we used the GnRH-secreting GT1-7 immortalized mouse cell line, a model system that has been extensively used in the past to study GnRH neuron biology (Mellon et al., 1990; Wetsel, 1995; Wierman et al., 1995). More importantly, GT1-7 hypothalamic neurons are responsive to androgen signaling because they express classical nuclear ARs

(Belsham et al., 1998; Brayman, Pepa, Berdy, et al., 2012; Brayman, Pepa, & Mellon, 2012). In addition, we showed that GT1-7 neurons express significant levels of Fgf8 mRNA, while our previous studies showed that GT1-7 neurons express Fgfr1 and Fgfr3 mRNA (Mott et al., 2010).

These cellular traits made GT1-7 neurons a suitable model system to study how androgen regulates Fgf8 transcription at the molecular level in GnRH secreting cells. We first examined two questions: 1) Does AR interact with the 5’ UTR promoter region of the mouse Fgf8 gene in

GT1-7 neurons, and 2) Does androgen modulate Fgf8 mRNA levels in GT1-7 neurons and in embryonic mouse OP cells? In addition, because the Fgf8 gene is enriched for CpG islands, we asked whether changes in DNA methylation affect Fgf8 mRNA levels in GT1-7 neurons.

Here we will discuss our data demonstrating that while AR interacts with the Fgf8 promoter region, this interaction was androgen-independent, and that androgen-treatment did not affect Fgf8 mRNA levels. In contrast, inhibition of DNA methyltransferases (DNMT) significantly upregulated Fgf8 mRNA levels.

MATERIALS AND METHODS

Timed-breeding of mice

Adult 129P2/OlaHsd*CD-1 male x female mice were timed-bred in the late afternoon in our animal facility (12L:12D cycle) with access to food and water ad libitum. All procedures were

41 approved by the Institutional Animal Care and Use Committee at Kent State University. In the morning, females with a sperm plug were denoted as embryonic day (E) 0.5.

Androgen response elements in the 5’UTR region of Fgf8

Identification and prediction of transcription factor consensus sites were determined through

MatInspector (Genomatix Software GmbH, Munich, Germany), which indicated the presence of three androgen response element (ARE) sites with a matrix similarity of greater than 0.88 (1 indicating a perfect match) within the 5000 bp 5’promoter region upstream from the transcriptional start site (Fig. 2.1.). The 5000 bp 5’ UTR of Fgf8 was selected due to its close proximity to the first coding exon in the Fgf8 gene and the presence of a large CpG island, which is well correlated to promoter regions (Calo & Wysocka, 2013). Given that the 5’ UTR has no

TATA box, the transcriptional start site(s) are currently unknown (Gemel et al., 1999). Previous literature has also indicated that the human Fgf8 5’UTR promoter region lies within this 5 Kb stretch (Gnanapragasam et al., 2002).

Figure 2.1. A) Immunocytochemistry of GnRH expressing GT1-7 neurons (gray) and B) negative control (no primary antibody). Scale bar = 25 µm. C) PCR in GT1-7 neurons. lane 1 = no cDNA, lane 2 = AR, lane 3 = Fgf8, lane 4= FgfR1, lane 5= GnRH. D) Analysis of the mouse

Fgf8 5’-UTR region (5 Kb upstream of the transcription start site (+1)) with Matinspector

(Genomatix) revealed the presence of three ARE sites upstream of the transcription start site with a shared consensus sequence. Relative position of F (forward) and R (reverse) flanking primer pairs (1, 2, 3) used to detect ARE 1 (-3077/-3058), ARE 2 (-1941/-1922), and ARE 3 (-528/-

509).

GT1-7 Neurons

Immortalized mouse GnRH neurons (GT1-7) (Generously donated by Dr. Pamela Mellon,

University of San Diego, CA) were grown in phenol-red free DMEM containing 4.5 g/L pyruvate and 548 mg/L-glutamine, 10% fetal bovine serum (ATCC, Manassas, Virginia), 1% pen/strep (Gemini Bio –Products, Sacramento, California), and 25 µg/ml Plasmocin (InvivoGen,

San Diego, California) (Mellon et al., 1990). Cells were kept in a humidified incubator at 37°C with 5% CO2. For our experiments, GT1-7 neurons were grown to 70-80% confluency, washed three times in a phosphate buffer solution and cultured for 16 hours in DMEM containing 10% dextran-charcoal stripped fetal bovine serum prior to pharmacological treatments (see below).

Immunohistochemistry in GT1-7 neurons

GT1-7 neurons were fixed with fresh ice-cold 4% paraformaldehyde for 30 minutes. Following, cells were washed in TBS (3 x 5 minutes) on a 2D rotator, incubated in primary rabbit polyclonal anti-GnRH (1:5,000, generously donated by Dr. Pei-San Tsai, University of Colorado Boulder,

CO) made in TBS/0.3% Triton-X (Fisher Scientific, Pittsburgh, PA) and 2% normal goat serum for 2 days at 4C. Cells were washed with TBS and incubated with biotinylated-goat anti-rabbit

(1:600) for 2 hours at room temperature followed by ABC (1:800) (Vector Laboratories,

Burlingame, CA) in TBS for 2 hours at room temperature, and reacted with 0.05% diaminobenzidine (Sigma-Aldrich, St. Louis, MO) containing 0.01% H2O2 in TBS for 20 minutes.

AR chromatin immunoprecipitation in GT1-7 neurons

Chromatin immunoprecipitation (ChIP) was used to examine whether AR interacted with the

5’UTR promoter region of the mouse Fgf8 gene in GT1-7 neurons in the presence or absence of dihydrotestosterone (DHT). GT1-7 cells were treated with vehicle (0.005% ethanol) or 100 nM

DHT for 4 hours. Following, GT1-7 neurons were processed for EZ-Magna-ChIP assays

(Millipore, Billerica, MA, USA) according to manufacturer’s instructions. Briefly, cells were cross-linked with 1% formalin for 15 min and lysed. The protein cross-linked genomic DNA was fragmented to 200-600 base pairs through sonication. Following, the fragments were immunoprecipitated using 1 µg of either rabbit polyclonal antibody against AR (Millipore,

Billerica, MA, USA) or control IgG (Millipore, Billerica, MA, USA) for two hours at 4ºC, which was pulled down using agarose-A beads coupled to magnets. Proteinase K (10 mg/ml) was used to reverse crosslinking, and DNA was isolated. The relative amount of AR occupancy on the identified ARE sites in the 5’UTR region of Fgf8 was measured using a Mastercycler EP

Realplex2 (Eppendorf, Hauppauge, NY) with SYBR Green PCR Master Mix (Roche, Basel,

Switzerland). For this purpose, three primer sets were designed to flank the identified ARE 1,

ARE 2 and ARE 3 sequences (Table 1). ChIP signal was normalized to background by using a

1% input, adjusted to 100%. Non-specific primers flanking upstream of the 5’ UTR region of

Fgf8 were used as negative controls for each pull-down. All results were performed in at least three independent experiments.

Table 2.1. Primer sequences used for detecting Fgf8, Fgfr1, GnRH and AR mRNA expression, and ARE containing DNA sequences within the Fgf8 promoter region.

Quantification of Fgf8 and Fgfr1 mRNA levels in DHT-treated GT1-7 neurons

GT1-7 cells were treated with vehicle (0.005% ethanol) or 100 nM DHT for 2, 4 or 48 hours.

This DHT concentration is a saturating amount given that the equilibrium dissociation constant

(Kd) of AR is ~2 nM (Wilson & French, 1976), and was used in previous studies investigating androgen signaling in GT1-7 neurons (Belsham et al., 1998). Additionally, transfection of human

AR-GFP plasmid into GT1-7 cells in the presence of 100 nM DHT was sufficient to translocate

AR into the nucleus (unpublished data). Total cellular RNA was extracted with TriPure (Roche,

Indianapolis, IN) according to the manufacturer’s instructions. RNA purity and concentration were measured using the Synergy H2 multi-mode reader with a Take3 Micro-Volume plate adapter (Biotek, VT). ProtoScript® II First Strand cDNA Synthesis Kit (New England Biolabs,

MA) was used to reverse transcribe 1 µg of total RNA according to the manufacturer’s instructions. RT-qPCR was performed in triplicate with gene-specific primers (Table 1) using a

Mastercycler EP Realplex2 (Eppendorf, Hauppauge, NY) with SYBR Green PCR Master Mix

(Roche, Basel, Switzerland). Relative Fgf8 and Fgfr1 mRNA expression levels were calculated using the ΔΔ-2CT method (Livak & Schmittgen, 2001). Hypoxanthine phosphoribosyltransferase

1 (Hprt-1) was used as a housekeeping gene.

Quantification Fgf8 and GnRH mRNA levels in E10.5 nasal explants treated with DHT

Adult pregnant wildtype female 129P2/OlaHsd*CD-1 mice were sacrificed at E10.5. The uterine horns were quickly removed from the mice and kept in sterile ice-cold phosphate-buffered saline

(Sigma-Aldrich, MO). Following, the nasal region containing the OP was surgically isolated using a dissection macroscope, placed on 0.85 µm Durapore membrane filters (Millipore,

Darmstadt, Germany), transferred to cell tissue culture inserts (Greiner Bio-One, Kremsmünster,

Austria ), and grown using the liquid-air interphase method with in phenol-red free Dulbecco's modified Eagle's medium (DMEM)/F12/glutamax (Thermofisher Scientific, MA) supplemented with B27 (Thermofisher Scientific, MA) and 1% pen/strep/myc (Sigma-Aldrich, MO) media.

E10.5 nasal explant tissues were collected, immediately flash frozen and kept at -80C and used as 0 days in vitro (DIV) samples or grown in culture for 3 DIV with media containing vehicle

(0.005% ethanol) or 100 nM DHT. Afterwards, total RNA was isolated from all explants (i.e., 0

DIV and 3 DIV). Relative Fgf8, Fgfr1, GnRH and AR mRNA expression levels were measured and calculated as described above.

Quantification Fgf8 and GnRH mRNA levels in DNA methyltransferase inhibitor-treated GT1-7 neurons

First, GT1-7 neurons were treated with vehicle (0.005% DMSO) or 1 µM 5-azacitidine (AZA) for 72 hours. Our AZA dose and length of treatment was based on previous studies showing that these conditions were able to induce gene expression (Y.-J. Xin et al., 2015; Zhou & Hu, 2015).

These studies were followed-up in second experiment where GT1-7 neurons were treated with

AZA for 72 hours in the presence or absence of 100 nM DHT for 4 hours. Total cellular RNA extraction and cDNA synthesis were performed (see above). Relative Fgf8 and GnRH mRNA expression levels were measured and calculated as described above.

Statistical analysis

Data were analyzed using Student t-tests or one-way analysis of variance (ANOVA) with treatment and/or DIV as between subject variables. Holm-Sidak tests were used for post hoc

48 analysis. Group n’s are given in the figure legends. Differences were considered significant if p <

0.05.

RESULTS

The mouse Fgf8 5’UTR regions harbors 3 ARE consensus sites.

Sequence analysis revealed three ARE sequences from the transcription start site (+1): 5’- actaaacatgttGTCCcag-3’ (ARE 1: -3077 bp), 5’-tgctttctctgtGTGCttg-3’ (ARE 2: -1941 bp), and

5’-agctgccgtcctGTCCttc-3’ (ARE 3: -528 bp) (Fig. 2.1.D).

GnRH, AR, Fgfr1 and Fgf8 mRNA expression in GT1-7 neurons

In these parametric studies, we established that our GT1-7 neurons were neurochemically similar to those used in other laboratories (Belsham et al., 1998; Brayman, Pepa, & Mellon, 2012;

Mellon et al., 1990; Mott et al., 2010). First, using ICC, we confirmed the presence of GnRH peptide in these cells (Fig. 2.1. A-B) Next, using PCR, we verified that our GT1-7 neurons express significant levels of GnRH, AR and Fgfr1 and Fgf8 mRNA (Fig. 2.1. C).

AR interacts with the 5’UTR region of Fgf8 in GT1-7 neurons

We then asked whether ARs are associated with the 5’UTR region of Fgf8 in GT1-7 neurons.

First, Student t-tests showed that ChIP with an AR-specific antibody was able to pull-down more

5’UTR Fgf8 DNA sequences that contain ARE 1, ARE 2 and ARE 3, in contrast to control IgG

(p < 0.05) (Fig. 2.2. A-C). In addition, Student t-tests showed that AR interacted with the AREs in a ligand-independent fashion indicating that DHT was unable to affect this interaction (Fig.

2.2. D-F).

Figure 2.2. ChIP gel electrophoresis in vehicle (VEH; n = 3) or DHT (n = 3) treated GT1-7 neurons (A-C) and RT-PCR analysis of ARE 1-3 sites (D-F). Using an AR-specific antibody, we found that AR interacts with the ARE 1-3 sites as compared to non-specific IgG immunoprecipitation (* indicates p < 0.05). However, DHT-treatment did not enhance AR binding to ARE 1-3 sites.

DHT did not affect Fgf8 mRNA levels in GT1-7 neurons

Previous studies showed that androgen was able to induce the mRNA expression levels of Fgf8 in mouse SC-3 breast and human LNCaP prostate cancer cells (Gnanapragasam et al., 2002;

Tanaka et al., 2002). In order to study whether that was also the case in GnRH-secreting neurons, we treated GT1-7 neurons with 100 nM DHT for 2, 4 or 48 hours. Our results showed that neither Fgf8 (Fig. 2.3. A-C) nor Fgfr1 (Fig. 2.3. D-F) mRNA levels were affected by DHT when compared to vehicle-treated GT1-7 neurons.

Figure 2.3. RT-qPCR for Fgf8 (A-C) and Fgfr1 (D-F) mRNA in GT1-7 neurons treated for 2 hrs with vehicle (n=4) or 100 nM DHT (n=4), 4 hrs with vehicle (n = 6) or 100 nM DHT (n = 6), or

48 hrs with vehicle (n = 3) or 100 nM DHT (n = 3). DHT treatment did not significantly alter

Fgf8 or Fgfr1 mRNA expression in GT1-7 neurons.

DHT did not affect Fgf8 mRNA levels in nasal explants

The inability of DHT to modulate Fgf8 mRNA expression in GT1-7 neurons may have been due to their immortalized status. Therefore, we used E10.5 nasal explants to investigate whether 3

DIV DHT treatment would be able to increase or decrease Fgf8, Fgfr1, AR or GnRH mRNA expression in an organotypic model system. Student t-tests showed that Fgf8 mRNA levels in 3

DIV nasal explants did not differ between vehicle-treated and DHT-treated 3 DIV nasal explant tissue (Fig. 2.4. A). Additional analysis showed that Fgf8 mRNA levels were, however, significantly lower in 3 DIV nasal explants as compared to 0 DIV nasal tissue (p < 0.05). Student t-tests showed that Fgfr1 mRNA levels in 3 DIV nasal explants did not differ between vehicle- treated and DHT-treated 3 DIV nasal explants (Fig. 2.4. B), although Fgfr1 mRNA levels were significantly higher in 3 DIV nasal explants as compared to 0 DIV nasal tissue (p < 0.05).

Student t-tests showed that GnRH mRNA levels in 3 DIV nasal explants did not differ between vehicle-treated and DHT-treated 3 DIV nasal explants tissue (Fig. 2.4. C). As expected, GnRH mRNA levels were significantly higher in 3 DIV nasal explants compared to 0 DIV nasal tissue

(p < 0.001). In contrast, Student t-tests showed that AR mRNA levels did not differ between vehicle-treated and DHT-treated 3 DIV nasal explants (Fig. 2.4. D). On average AR mRNA levels seem to be higher 0 DIV nasal tissue compared to 3 DIV nasal explants, but however, did not reach significance.

Figure 2.4. RT-qPCR for Fgf8, Fgfr1, GnRH, AR mRNA nasal explants treated for 3 DIV with vehicle (n = 3) or 100 nM DHT (n = 3 animals) (A-D). Inset panels represent comparisons between 0 DIV (n = 3) vs collapsed 3 DIV nasal explant expression data (n = 6), which showed that Fgf8 mRNA significantly decreased (A), Fgfr1 mRNA significantly increased (B), GnRH mRNA significantly increased (C), and no significant difference in AR mRNA expression (D) after 3 DIV. * indicates p < 0.05.

Inhibition of DNA methyltransferase upregulated Fgf8 mRNA levels in GT1-7 neurons

Here, we asked whether changes in DNA methylation affect Fgf8 mRNA levels in GT1-7 neurons, because the Fgf8 gene structure is enriched with CpG islands. In our first study, Student t-tests showed that compared to vehicle treatment, AZA treatment significantly increased Fgf8 (p

< 0.05; Fig. 2.5. A) mRNA levels. In contrast, AZA dramatically reduced GnRH mRNA levels

(p < 0.0001; Fig. 2. 5 B). In our second study, one-way ANOVA showed a significant treatment effect on Fgf8 mRNA levels (F(3, 15) = 10.5, p < 0.005; Fig. 2.5. C). Post hoc analysis showed that AZA significantly increased Fgf8 mRNA levels compared to vehicle (p < 0.01), DHT (p <

0.005) and AZA + DHT (p < 0.005) treatment groups. Interestingly, while we confirmed our earlier results indicating that DHT does not affect Fgf8 mRNA levels, we found that DHT prevented AZA-dependent increase in Fgf8 mRNA levels. Moreover, one-way ANOVA revealed a significant treatment effect on GnRH mRNA levels (F(3, 15) = 13.7, p < 0.005; Fig. 2.5. D).

Post hoc analysis showed that compared to vehicle and DHT treatment groups, AZA treatment significantly reduced GnRH mRNA levels (p < 0.05). Similarly, compared to vehicle and DHT treatment groups, GnRH mRNA levels were significantly lower in GT1-7 neurons treated with

AZA + DHT (p < 0.005). Moreover, compared to vehicle-treatment, DHT treatment did not affect GnRH mRNA levels, nor was DHT able to prevent AZA-dependent downregulation of

GnRH mRNA levels.

Figure 2.5. RT-qPCR for Fgf8 and GnRH mRNA in GT1-7 neurons treated with vehicle (n = 4) or 1 µM AZA (n = 4) for 72 hours. AZA-treatment significantly increased Fgf8 mRNA expression (p < 0.05) (A) and reduced GnRH mRNA levels (p < 0.0001) (B). RT-qPCR for Fgf8 and GnRH mRNA expression in GT1-7 neurons treated with vehicle or AZA for 3 days and vehicle or 100 nM DHT for 4 hours (C-D). One-way ANOVA showed a significant treatment effect on Fgf8 mRNA levels (F(3, 15) = 10.5, p < 0.005) (C). Post hoc analysis showed that

AZA increased Fgf8 mRNA levels compared to vehicle (n = 4; p < 0.01), DHT (n = 4; p < 0.005) and AZA + DHT (n = 4; p < 0.005) treatment groups. One-way ANOVA showed a significant treatment effect on GnRH mRNA levels (F(3, 15) = 13.7, p < 0.005) (D). Post hoc analysis showed that compared to vehicle and DHT treatment groups, AZA treatment significantly reduced GnRH mRNA levels (n = 4; p < 0.05). GnRH mRNA levels were significantly lower in

GT1-7 neurons treated with AZA + DHT (n = 4; p < 0.005). DHT treatment did not affect GnRH mRNA levels, nor was DHT able to prevent AZA-dependent downregulation of GnRH mRNA levels. * or different letter indicate p < 0.05.

DISCUSSION

Several novel discoveries are reported in this study. First, we showed that AR interacts with 3 specific ARE sites (ARE 1: -3077 bp; ARE 2: -1941 bp; ARE 3: -528 bp) in the 5’UTR region of the mouse Fgf8 gene in GT1-7 neurons. Furthermore, we found that unliganded AR was already recruited to all 3 ARE containing 5’UTR regions, which our data found to be unaffected by the presence of DHT. Second, in contrast to our hypothesis, which was based on previous studies in

SC3 and LNCaP cells, DHT did not modulate Fgf8 mRNA levels in GT1-7 neurons or nasal explants. In contrast, inhibition of DNMT using AZA significantly upregulated Fgf8 mRNA levels, which concomitant DHT treatment prevented. This suggests that although the 5’UTR region of the Fgf8 gene allows for AR binding on three separate ARE consensus sites, the function of androgen signaling may not be to upregulate Fgf8 transcription, but rather to moderate/inhibit the upregulatory effects of other molecular processes, such as DNA methylation.

ChIP assays showed that AR interacts with the Fgf8 promoter, which suggests that the

ARE consensus sites found in the 5’UTR of Fgf8 are able to enlist ARs in GT1-7 neurons. These results follow earlier studies reporting that ARs were closely associated with the human Fgf8 promoter region LNCaP cells (Gnanapragasam et al., 2002). However, to our surprise, our DHT treatments did not increase AR recruitment to the mouse Fgf8 promoter region. Contrarily, earlier findings in LNCaP cells showed that the synthetic non-aromatizable androgen, mibolerone was able to cause enhanced AR recruitment to the human Fgf8 promoter region

(Gnanapragasam et al., 2002). Specifically, these earlier studies showed that mibolerone- dependent AR recruitment occurred only at the most distal ARE consensus site (out of 3 AREs) on the human Fgf8 promoter region (Gnanapragasam et al., 2002). This discrepancy may be due

57 to technical differences in detection approaches. In our studies, qPCR was used to quantify AR recruitment to the mouse Fgf8 promoter, whereas AR recruitment to the human Fgf8 promoter in

LNCaP cells was quantified after visualizing the PCR product on film. Alternatively, and more likely explanation for these partially contradictory results may be due to inherent biochemical differences between the androgenic agonists: DHT is an endogenous and naturally occurring androgen, whereas mibolerone is synthetic (Gnanapragasam et al., 2002). Indeed, recent studies showed that although DHT and mibolerone are non-aromatizable androgens, mibolerone may have distinct non-physiological effects on the cell's molecular signaling machinery, such as that mibolerone can activate progestin receptors in breast cancer cells, a non-physiological effect that

DHT does not exhibit (Cops et al., 2008). Similarly, the expression of some miRNAs in LNCaP cells, such as miR-120 can be induced in the presence of mibolerone, but not DHT (Segal et al.,

2015). Taken together, we hypothesize that AR recruitment to the Fgf8 promoter under normal physiological conditions is independent of the cellular and extracellular androgen milieu, however a more extensive analysis of AR/ARE function are needed to definitively rule out an androgen-mediated effect on Fgf8 expression.

The lack of DHT-dependent induction of Fgf8 mRNA in GT1-7 neurons was unexpected, especially in light the data from earlier studies in SC3 and LNCaP cells (Gnanapragasam et al.,

2002; Ohuchi et al., 1994; Tanaka et al., 2002; Yamanishi et al., 1995). The ability of AR to drive transcription depends on the recruitment of specific co-regulators, which may act directly or indirectly with AR to modulate its ability to activate gene transcription (Chung & Auger,

2013). It is well known that in addition to the presence of its cognate androgenic ligand, the cellular complement of co-factors, such as SRC-1 and ARA-70 plays a role in AR’s ability to drive transcription, and that this complement is cell type dependent. Therefore, it is possible that

GT1-7 neurons do not have the pre-requisite complement of co-factors that allows DHT to appropriately activate AR to induce Fgf8 mRNA levels as it is the case in SC3 and LNCaP cells.

Alternatively, regulation of Fgf8 transcription may be under the control of DNA methylation status (Chung & Auger, 2013), an inference that is supported by our results. In addition to harboring multiple CpG islands upstream and downstream of the transcription start site (http://www.ncbi.nlm.nih.gov/epigenomics/), we found that AZA induced Fgf8 mRNA levels in GT1-7 neurons. Although we currently do not known whether AZA directly changed the DNA methylation status of the Fgf8 promoter, our results follow previous studies indicating that hypomethylation in primary rhabdomyosarcoma tumors was correlated with higher Fgfr1 mRNA expression levels in (Goldstein et al., 2007). These studies confirmed earlier experiments in which DHT treatment for 4 hours does not affect Fgf8 mRNA levels; however, to our surprise,

DHT did eliminate the AZA-dependent rise in Fgf8 mRNA levels. We speculate that in contrast to our original hypothesis, androgen signaling may act to act downregulate DNA methylation- dependent upregulation of Fgf8 mRNA expression, and hence limit the morphogenetic or proliferative effects of FGF8 signaling (Ornitz & Itoh, 2015). Nonetheless, these results indicate that changes in DNA methylation status can induce Fgf8 transcription, and that androgen signaling is functional.

In contrast, AZA caused a dramatic decrease in GnRH mRNA levels in GT1-7 neurons.

This result was unexpected, because evidence from earlier studies in Rhesus monkey nasal explants showed that demethylation of the GnRH promoter correlates with a rise in GnRH mRNA levels (Kurian et al., 2010). This discrepancy may be due to the inherent differences between our experimental model systems. Alternatively, and more likely, the decrease in GnRH mRNA levels may be related to the increase in Fgf8 mRNA expression. Indeed, previous studies

59 showed that treatment with recombinant FGF8b protein not only reduced GnRH promoter activity, but also GnRH mRNA levels in GT1-7 neurons (Mott et al., 2010). Currently, we are studying whether FGF8 signaling overrode the rise of GnRH mRNA levels due to the DNA demethylating effects of AZA.

In conclusion, our data showed that while AR interacts with the Fgf8 promoter region, this interaction was androgen-independent, and that androgen treatment alone did not affect

Fgf8 mRNA levels, indicating that androgen signaling does not induce Fgf8 transcription in

GT1-7 GnRH-secreting neurons or nasal explant cells. In contrast, our studies did show that inhibition of DNA methyltransferases significantly upregulated Fgf8 mRNA levels indicating that Fgf8 transcriptional activity may be dependent on methylation status.

REFERENCES

Belsham, D. D., Evangelou, A., Roy, D., Duc, V. L., & Brown, T. J. (1998). Regulation of

gonadotropin-releasing hormone (GnRH) gene expression by 5alpha-dihydrotestosterone

in GnRH-secreting GT1-7 hypothalamic neurons. Endocrinology, 139(3), 1108-1114.

doi: 10.1210/endo.139.3.5846

Brayman, M. J., Pepa, P. A., Berdy, S. E., & Mellon, P. L. (2012). Androgen receptor repression

of GnRH gene transcription. Mol Endocrinol, 26(1), 2-13. doi: 10.1210/me.2011-1015

Brayman, M. J., Pepa, P. A., & Mellon, P. L. (2012). Androgen receptor repression of

gonadotropin-releasing hormone gene transcription via enhancer 1. Mol Cell Endocrinol,

363(1-2), 92-99. doi: 10.1016/j.mce.2012.07.012

Calo, E., & Wysocka, J. (2013). Modification of enhancer chromatin: what, how, and why? Mol

Cell, 49(5), 825-837. doi: 10.1016/j.molcel.2013.01.038

Chung, W. C., & Auger, A. P. (2013). Gender differences in neurodevelopment and epigenetics.

Pflugers Arch, 465(5), 573-584. doi: 10.1007/s00424-013-1258-4

Chung, W. C., Moyle, S. S., & Tsai, P. S. (2008). Fibroblast growth factor 8 signaling through

fibroblast growth factor receptor 1 is required for the emergence of gonadotropin-

releasing hormone neurons. Endocrinology, 149(10), 4997-5003. doi: 10.1210/en.2007-

1634

Cops, E. J., Bianco-Miotto, T., Moore, N. L., Clarke, C. L., Birrell, S. N., Butler, L. M., &

Tilley, W. D. (2008). Antiproliferative actions of the synthetic androgen, mibolerone, in

breast cancer cells are mediated by both androgen and progesterone receptors. J Steroid

Biochem Mol Biol, 110(3-5), 236-243. doi: 10.1016/j.jsbmb.2007.10.014

Crossley, P. H., & Martin, G. R. (1995). The mouse Fgf8 gene encodes a family of polypeptides

and is expressed in regions that direct outgrowth and patterning in the developing

embryo. Development, 121(2), 439-451.

Evans, R. M. (1988). The steroid and thyroid hormone receptor superfamily. Science, 240(4854),

889-895. doi: 10.1126/science.3283939

Gemel, J., Jacobsen, C., & MacArthur, C. A. (1999). Fibroblast growth factor-8 expression is

regulated by intronic engrailed and Pbx1-binding sites. J Biol Chem, 274(9), 6020-6026.

Gnanapragasam, V. J., Robson, C. N., Neal, D. E., & Leung, H. Y. (2002). Regulation of FGF8

expression by the androgen receptor in human prostate cancer. Oncogene, 21(33), 5069-

5080. doi: 10.1038/sj.onc.1205663

Goldstein, M., Meller, I., & Orr-Urtreger, A. (2007). FGFR1 over-expression in primary

rhabdomyosarcoma tumors is associated with hypomethylation of a 5' CpG island and

abnormal expression of the AKT1, NOG, and BMP4 genes. Genes Chromosomes

Cancer, 46(11), 1028-1038. doi: 10.1002/gcc.20489

Kawauchi, S., Beites, C. L., Crocker, C. E., Wu, H. H., Bonnin, A., Murray, R., & Calof, A. L.

(2004). Molecular signals regulating proliferation of stem and progenitor cells in mouse

olfactory epithelium. Dev Neurosci, 26(2-4), 166-180. doi: 10.1159/000082135

Kurian, J. R., Keen, K. L., & Terasawa, E. (2010). Epigenetic changes coincide with in vitro

primate GnRH neuronal maturation. Endocrinology, 151(11), 5359-5368. doi:

10.1210/en.2010-0555

Livak, K. J., & Schmittgen, T. D. (2001). Analysis of relative gene expression data using real-

time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods, 25(4), 402-408.

doi: 10.1006/meth.2001.1262

Mellon, P. L., Windle, J. J., Goldsmith, P. C., Padula, C. A., Roberts, J. L., & Weiner, R. I.

(1990). Immortalization of hypothalamic GnRH neurons by genetically targeted

tumorigenesis. Neuron, 5(1), 1-10.

Meyers, E. N., Lewandoski, M., & Martin, G. R. (1998). An Fgf8 mutant allelic series generated

by Cre- and Flp-mediated recombination. Nat Genet, 18(2), 136-141. doi:

10.1038/ng0298-136

Mott, N. N., Chung, W. C., Tsai, P. S., & Pak, T. R. (2010). Differential fibroblast growth factor

8 (FGF8)-mediated autoregulation of its cognate receptors, Fgfr1 and Fgfr3, in neuronal

cell lines. PLoS One, 5(4), e10143. doi: 10.1371/journal.pone.0010143

Ohuchi, H., Yoshioka, H., Tanaka, A., Kawakami, Y., Nohno, T., & Noji, S. (1994).

Involvement of androgen-induced growth factor (FGF-8) gene in mouse embryogenesis

and morphogenesis. Biochem Biophys Res Commun, 204(2), 882-888. doi:

10.1006/bbrc.1994.2542

Ornitz, D. M., & Itoh, N. (2015). The Fibroblast Growth Factor signaling pathway. Wiley

Interdiscip Rev Dev Biol, 4(3), 215-266. doi: 10.1002/wdev.176

Sabado, V., Barraud, P., Baker, C. V., & Streit, A. (2012). Specification of GnRH-1 neurons by

antagonistic FGF and retinoic acid signaling. Dev Biol, 362(2), 254-262. doi:

10.1016/j.ydbio.2011.12.016

Schwanzel-Fukuda, M., & Pfaff, D. W. (1989). Origin of luteinizing hormone-releasing hormone

neurons. Nature, 338(6211), 161-164. doi: 10.1038/338161a0

Segal, C. V., Koufaris, C., Powell, C., & Gooderham, N. J. (2015). Effects of treatment with

androgen receptor ligands on microRNA expression of prostate cancer cells. Toxicology,

333, 45-52. doi: http://dx.doi.org/10.1016/j.tox.2015.04.002

Sun, X., Meyers, E. N., Lewandoski, M., & Martin, G. R. (1999). Targeted disruption of Fgf8

causes failure of cell migration in the gastrulating mouse embryo. Genes Dev, 13(14),

1834-1846.

Tanaka, A., Kamiakito, T., Takayashiki, N., Sakurai, S., & Saito, K. (2002). Fibroblast growth

factor 8 expression in breast carcinoma: associations with androgen receptor and

prostate-specific antigen expressions. Virchows Arch, 441(4), 380-384. doi:

10.1007/s00428-002-0687-9

Tanaka, A., Miyamoto, K., Minamino, N., Takeda, M., Sato, B., Matsuo, H., & Matsumoto, K.

(1992). Cloning and characterization of an androgen-induced growth factor essential for

the androgen-dependent growth of mouse mammary carcinoma cells. Proc Natl Acad Sci

U S A, 89(19), 8928-8932. doi: 10.1073/pnas.89.19.8928

Tsai, P. S., Brooks, L. R., Rochester, J. R., Kavanaugh, S. I., & Chung, W. C. (2011). Fibroblast

growth factor signaling in the developing neuroendocrine hypothalamus. Front

Neuroendocrinol, 32(1), 95-107. doi: 10.1016/j.yfrne.2010.11.002

Wetsel, W. C. (1995). Immortalized hypothalamic luteinizing hormone-releasing hormone

(LHRH) neurons: a new tool for dissecting the molecular and cellular basis of LHRH

physiology. Cell Mol Neurobiol, 15(1), 43-78.

Wierman, M. E., Bruder, J. M., & Kepa, J. K. (1995). Regulation of gonadotropin-releasing

hormone (GnRH) gene expression in hypothalamic neuronal cells. Cell Mol Neurobiol,

15(1), 79-88.

Wilson, E. M., & French, F. S. (1976). Binding properties of androgen receptors. Evidence for

identical receptors in rat testis, epididymis, and prostate. J Biol Chem, 251(18), 5620-

5629.

Wray, S., Grant, P., & Gainer, H. (1989). Evidence that cells expressing luteinizing hormone-

releasing hormone mRNA in the mouse are derived from progenitor cells in the olfactory

placode. Proc Natl Acad Sci U S A, 86(20), 8132-8136.

Xin, Y.-J., Yuan, B., Yu, B., Wang, Y.-Q., Wu, J.-J., Zhou, W.-H., & Qiu, Z. (2015). Tet1-

mediated DNA demethylation regulates neuronal cell death induced by oxidative stress.

Scientific Reports, 5, 7645. doi: 10.1038/srep07645

Yamanishi, H., Tanaka, A., Miyamoto, K., Takeda, M., Nishizawa, Y., Koga, M., . . .

Matsumoto, K. (1995). An essential role of androgen-induced growth factor in

glucocorticoid-dependent autocrine loop in Shionogi carcinoma 115 cells. J Steroid

Biochem Mol Biol, 52(1), 49-53. doi: 10.1016/0960-0760(94)00148-f

Zhou, Y., & Hu, Z. (2015). Genome-wide demethylation by 5-aza-2'-deoxycytidine alters the

cell fate of stem/progenitor cells. Stem Cell Rev, 11(1), 87-95. doi: 10.1007/s12015-014-

9542-z

CHAPTER III THE INDIRECT ROLE OF RETINOIC ACID IN THE DEVELOPING OLFACTORY PLACODE.

INTRODUCTION

Patterning of the embryonic nervous system is coordinated, in part, by diffusible morphogen gradients which outline diverse cell fate specifications and coordinate transcription factor networks. In addition to FGF8 signaling in the olfactory placode

(OP), retinoic acid (RA) signaling is a well-known contributor for establishing cell fate

(Cho et al., 1998; Cho et al., 2001; Huang et al., 1985; Sabado et al., 2012). Retinoic acid is derivative of vitamin A, which acts through the nuclear receptors, retinoic acid receptor

(RAR) and retinoic X receptor (RXR) to regulate gene transcription (Evans &

Mangelsdorf, 2014; Giguere et al., 1987; Petkovich et al., 1987). RAR receptors dimerize and bind to cis-acting retinoic acid response elements (RARE) found in gene regulatory regions on DNA. Like other nuclear receptors, RARs can recruit co- activator and co- repressor complexes to exert ligand-dependent functions, mainly by serving as docking sites for chromatin modifying enzymes. Circulating retinol is bound to retinol-binding protein 4 (RBP4), a carrier protein, and is then oxidized to retinaldehyde by cytosolic alcohol dehydrogenases (ADHs) or retinol dehydrogenases (RDHs) (Kanai et al., 1968).

Retinaldehyde is further oxidized by retinaldehyde dehydrogenases (RALDHs) which produce either all-trans retinoic acid (ATRA) or 9-cis retinoic acid (Niederreither et al.,

2002; Yoshida et al., 1998; Zhao et al., 1996) . Importantly, transgene reporters have

66 demonstrated that RALDH expression patters correlate with RA activity in tissue and provide a viable way of mapping RA activity (Ang et al., 1996; Niederreither et al., 1997).

During early embryonic development, RA is secreted from mesodermal tissue and is involved in the induction of neuronal differentiation, apoptosis, proliferation as well as retinal development (Das et al., 2014; Dubey et al., 2018; Mey et al., 1997; Pourjafar et al., 2017).

FGFs and WNTs are similarly secreted during early development and function in concert with

RA to generate distinct patterning signals. Interestingly, during patterning of vertebrate trunk development, RA and FGF signaling have opposing diffuse signaling patterns. In the developing embryonic day (E) 7.5 mouse embryo, RA is secreted in the presomitic mesoderm, and can indirectly influence trunk development by repressing Fgf8 expression during body axis extension

(Kumar & Duester, 2014). Duester et. al. showed that the promoter region of FGF8 contains a conserved RARE, which binds RARs and recruits repressive histone modifications to transcriptionally repress Fgf8 (Kumar & Duester, 2014). Retinoic acid has also been shown to restrict Fgf8 expression to the primitive ectoderm, preventing Fgf8 expression in the neuroectoderm (Kumar & Duester, 2014; Sirbu & Duester, 2006). In Raldh2-/- embryos, Fgf8 mRNA exhibits a shift in left-right symmetry from the epiblast into the node ectoderm, where

Fgf8 mRNA is not normally detected (Sirbu & Duester, 2006). Furthermore, the mutant Raldh2-/- embryos could be rescued with maternal retinoic acid supplements and restore normal Fgf8 expression (Sirbu & Duester, 2006). While these experiments are seminal in demonstrating the ability of RA to regulate Fgf8 transcription and ultimately, embryonic patterning, it is currently unknown if RA also regulates Fgf8 in the olfactory placode during GnRH neuron onset.

Early studies in hypothalamic (POA-MBH) fragments demonstrated that RA treatment was able to reduce GnRH mRNA (Cho et al., 1998). Conversely, in both GT1-1 neuronal lines

67 and hypothalamic fragments, RA increased GnRH peptide production (Cho et al., 1998).

Luciferase assays further demonstrated that GnRH promoter constructs around a putative RARE site could be activated by ATRA (Cho et al., 1998). These studies posit that RA transcriptionally represses GnRH transcription but, has an activating function on GnRH peptide release and promoter activity. Thus, it remains unclear what regulatory effect RA has on the function of

GnRH neurons. Similarly, these studies explored adult GnRH cell lines and hypothalamic fragments. Here, we asked whether RA could affect GnRH mRNA and neuronal migration during early GnRH development in GN11 cells and in OP cultures.

In this study, we hypothesized that RA causes RARs to bind to the RARE on the Fgf8 promoter during OP development and modulate the both expression of Fgf8 mRNA and therefore control the emergence of GnRH neurons from the olfactory placode. Here, we performed experiments to determine the molecular control of Fgf8 as well as GnRH transcription in the presence of RA or RA synthesis inhibitors. We also analyzed the functional effects of RA on GnRH neuron cell migration and proliferation using a cell scraping assay.

MATERIALS AND METHODS

Timed breeding of mice and nasal explant cultures

Adult wildtype 129P2/OlaHsd*CD-1 male x female mice were timed-bred in the late afternoon in our animal facility (12L:12D cycle) with access to food and water ad libitum. All procedures were approved by the Institutional Animal Care and Use Committee at Kent State University. In the morning, females with a sperm plug were denoted as embryonic day (E) 0.5. Adult pregnant female 129P2/OlaHsd*CD-1 mice were sacrificed at E9.5, E10.5 or E13.5. The mouse uterine horns were quickly removed and kept in sterile ice-cold phosphate-buffered saline (Sigma-

Aldrich, P3813). Following, the nasal region containing the OPs was surgically isolated using a dissection microscope, placed on 0.65 μm Durapore membrane filters (Millipore, DVPP04700), transferred to cell tissue culture inserts (Corning, 353095), and grown using the liquid-air interphase method with in phenol-red free Dulbecco's modified Eagle's medium

(DMEM)/F12/glutamax (Thermofisher Scientific, 10565018) supplemented with B27

(Thermofisher Scientific, 17504044) and 1% pen/strep/myc (Sigma-Aldrich, A5955) media. For mRNA expression assays, and Chromatin immunoprecipitation experiments, E9.5, E10.5, or

E13.5 nasal explant tissues were collected, immediately flash frozen and kept at -80C.

RA chromatin immunoprecipitation in GT1-7 neurons

ChIP was used to examine RARs and histone modifications (H3K4me3 and H3K27me3) along the promoter region of Fgf8. Embryos were dissected at E9.5, E10.5 or E13.5, and pooled in microcentrifuge tubes containing 4-6 OPs per sample. Chromatin was harvested using the EZ-

Magna-ChIP kit (Millipore, 17-10086) according to manufacturer's instructions. Briefly, cells were cross-linked with 1% formalin for 10 min and lysed. The protein cross-linked genomic

DNA was fragmented to 200–600 base pairs using sonication. All chromatin samples were verified for correct shearing density on an agarose gel before continuing. Magnetic Protein A/G beads were blocked with salmon sperm for 2 hours and washed in PBS. Following, the fragments were immunoprecipitated using 1.25 μg of either rabbit polyclonal antibody against pan-RAR

(Santa Cruz, M-454), H3K4me3, and H3K27me3 (Active Motif 39915, 39155) or control IgG

(Millipore, 12-370) for 3 hours at 4°C on a 3D rotator. Magnetic beads were pulled out of solution using a magnet, and bound fragments were washed four times for ten minutes.

Proteinase K (10 mg/ml) was used to reverse crosslinking, and DNA was isolated using phenol

69 chloroform. The relative amount of protein occupancy on the identified sites on the promoter of Fgf8 was measured using a Mastercycler EP Realplex2 (Eppendorf, EPPE6300000.604) with

SYBR Green PCR Master Mix (Roche, 04707516001). For this purpose, primer sets were designed to flank the RARE on the promoter region. Relative enrichment was calculated using the percent input method, where each immunoprecipitation is adjusted to input loading controls, and compared to negative control IgG for significant antibody enrichment. Non-specific primers flanking the 3’UTR of Fgf8, were also used in the ChIP.

Cell Culture

Immortalized mouse GnRH secreting GN11 neurons (generously donated by Dr. Sally Radovick,

Robert Wood Johnson Medical School, NJ) were grown in phenol-red free DMEM containing

4.5 g/L pyruvate and 548 mg/L-glutamine, 10% fetal bovine serum (ATCC, 30-2020), 1% pen/strep (Sigma-Aldrich, A5955) (Mellon et al., 1990). Cells were kept in a humidified incubator at 37°C with 5% CO2.

Pharmacological treatments

OP explants were treated for 72 h (i.e., 3 days in vitro (DIV)) in the presence of vehicle (0.005%

DMSO) or 1 μM all-trans retinoic acid (Tocris Biosciences, 3842) or 50 μM citral (Sigma-

Aldrich, C83007). Our ATRA dose and length of treatment was based on previous studies showing that these conditions were able to induce gene expression (Cho et al., 1998). Total cellular RNA extraction and cDNA synthesis were performed as described below.

RT-qPCR

Total cellular RNA was extracted with TriPure (Roche, 11667165001) according to the manufacturer’s instructions. RNA purity and concentration were measured using the Synergy H2 multi-mode reader with a Take3 Micro-Volume plate adapter (Biotek). ProtoScript® II First

Strand cDNA Synthesis Kit (New England Biolabs, E6560L). ProtoScript® II First Strand cDNA Synthesis Kit was used to reverse transcribe 0.5 μg of total RNA. RT-qPCR was performed in triplicate with gene-specific, intron-spanning primers using a Mastercycler EP

Realplex2 (Eppendorf, EPPE6300000.604) with SYBR Green PCR Master Mix (Roche,

04707516001). Relative mRNA expression levels were calculated using the ΔΔ-2CT method

(Livak & Schmittgen, 2001). Hypoxanthine phosphoribosyltransferase 1 (Hprt-1) was used as a housekeeping gene.

GnRH neuron cell scraping assays

Immortalized mouse GN11 neurons were grown in phenol-red free DMEM 1% pen/strep

(Sigma-Aldrich, A5955) in a humidified incubator at 37°C with 5% CO2 until ~60% confluency.

A cell scraper with a 11 mm silicone blade was used to scrape cells in the center of the plate.

Cells were briefly washed in sterile PBS, and media was replaced with either vehicle, 0.2 μM,

0.5 μM or 1 μM ATRA. Neuron migration/proliferation into the scraped area was measured for 3 days following treatment. Cell number was manually counted and recorded.

Quantification Fgf8 and GnRH mRNA levels in E10.5 nasal explants treated with RA

Adult pregnant wildtype female 129P2/OlaHsd*CD-1 mice were sacrificed at E10.5. The uterine horns were quickly removed from the mice and kept in sterile ice-cold phosphate-buffered saline

(Sigma-Aldrich, MO). Following, the nasal region containing the OP was surgically isolated

71 using a dissection macroscope, placed on 0.85 μm Durapore membrane filters (Millipore,

Darmstadt, Germany), transferred to cell tissue culture inserts (Greiner Bio-One, Kremsmünster,

E10.5 nasal explant tissues were collected, immediately flash frozen and kept at -80C and used as 0 days in vitro (DIV) samples or grown in culture for 3 DIV with media containing vehicle

(0.005% DMSO) or 1 µm all-trans retinoic acid. Afterwards, total RNA was isolated from all explants (i.e., 0 DIV and 3 DIV). Relative Fgf8, GnRH and RALDH mRNA expression levels were measured and calculated as described above.

Statistical analysis

Data were analyzed using Student t-tests or one-way analysis of variance (ANOVA) with treatment and/or DIV as between subject variables. Holm-Sidak tests were used for post hoc analysis. Group n’s are given in the figure legends. GN11 cell scraping assay was measured using a 2-way ANOVA with repeated measures. Differences were considered significant if p <

0.05.

RESULTS

Expression of retinal dehydrogenases in the OP

While studies in the developing mouse E7.5 neural tube demonstrate the presence of RALDH expression, whether these enzymes were present in the OP were unknown. Here, we demonstrated that both RALDH2 and RALDH3 were present in the mouse OP (Fig. 3.1.),

72 however, no detectable RALDH1 could be measured. Interestingly, we found a significant increase in RALDH2 expression during the emergence of GnRH neurons (p < 0.00001).

Figure 3.1. RALDH expression in the OP. A) RALDH2 and RALDH3 mRNA expression in

E10.5 or E13.5 mouse OP (n = 4). RALDH1 mRNA was not detectable. * indicates p < 0.00001;

Student’s t-test.

Fgf8 mRNA expression in ATRA-treated GN11 neurons

The increase in mRNA expression of RALDH2 in the developing OP suggested that RA may be an important factor in the development of GnRH neurons. Specifically, we hypothesized that RA signaling could decrease Fgf8 mRNA expression, which would in turn, affect the emergence of

GnRH neurons. To test the ability of RA to regulate Fgf8 transcription in GnRH neurons specifically, we treated immortalized GN11 neurons with ATRA for 16 hours and measured Fgf8 mRNA expression. We found that ATRA treatment did not affect Fgf8 mRNA expression.

Figure 3.2. Fgf8 expression in the GN11 cell line. A) Fgf8 mRNA expression in the presence of vehicle, 0.01, 0.1, or 1µM ATRA for 16 hours (n = 4).

We then hypothesized other cell types within the OP may be more responsive to RA signaling.

For these experiments we used an OP culture system to measure 3DIV of ATRA treatment.

Interestingly, we found that 1 µM ATRA could effectively reduce the amount of GnRH mRNA produced in the OP (p < 0.001). However, there was no significant difference in Fgf8 mRNA expression in ATRA treated OP cultures.

Figure 3.3. RA represses GnRH expression in the OP. A) RT-qPCR for Fgf8, and GnRH mRNA in vehicle vs 1 µM RA-treated E10.5 mouse OP explants (n = 4) for 3 DIV; Student’s t-test.

RA does not bind to the RARE in E9.5 or E13.5 OP cultures nor recruit repressive histone modifications

To test whether RA could directly bind the Fgf8 promoter region at a previously identified conserved RARE, we performed ChIP experiments at E9.5 and E13.5 in OP tissue. We found that as expected, RA did not bind to the promoter region of FGF8 at E9.5, when transcription is high. At E13.5, RA was also not significantly enriched at the RARE site, indicating that RA does not directly interact with the FGF8 locus in the OP at E9.5 or E13.5. In addition, previous studies found that RARs were able to recruit repressive chromatin modifications to modify gene expression. We concomitantly tested if the RARE sites on the Fgf8 promoter could recruit the repressive histone modification H3K27me3 (p <0.01). In the E13.5 embryo, we found that the

RARE site was enriched for H3K27me3, a repressive histone modification, but not H3K4me3, an active histone modification.

Figure 3.4. RARs do not bind to the Fgf8 promoter during OP development. Pan-RAR,

H3K4me3 or H3K27me3 ChIP-RT-qPCR of 3-6 pooled E9.5 or E13.5 mouse OPs on the RARE site of the Fgf8 promoter (n = 4). * indicates p < 0.01. Student’s t-test.

Expression of Fgf8 and GnRH mRNA in citral-treated OP cultures

We next tested whether citral, a RA inhibitor, could affect Fgf8 and GnRH mRNA transcription in OP cultures. We used 50 M citral, to treat OP cultures for 3DIV. Our findings suggested that inhibition of RA synthesis could effectively increase Fgf8 mRNA transcription (p < 0.05), indicating RA normally has a repressive role. However, in the presence of citral, GnRH mRNA was downregulated (p < 0.05), suggesting that RA synthesis is required for GnRH transcription in the early OP.

Figure 3.5. RALDH inhibition induces Fgf8 expression in the OP. A) RT-qPCR for Fgf8, GnRH mRNA in vehicle (n = 4) vs 50 µM citral-treated E10.5 mouse OP explants for 3 DIV * indicates p < 0.05; Student’s t-test.

RA treatment increases GnRH neuron migration and/or proliferation in-vitro

Because citral and ATRA appeared to have opposed regulatory functions on GnRH mRNA transcription, we asked whether RA had a functional role in GnRH neurons. To test this, we used migratory GN11 neurons to measure GnRH migration and/or proliferation. We hypothesized that since RA was able to decrease GnRH mRNA in the OP during emergence of the neurons, that

RA may affect either the migration of GnRH neurons into the hypothalamus or proliferation of the progenitor cell population. Here, we designed a plate scraping assay to test the ability of the neurons to migrate and/or proliferate in the presence of ATRA over the course of 3 days. A 2- way ANOVA with repeated measures was used to determine the effect of time and concentration. Our results indicated that ATRA could induce GnRH neuron migration and/or proliferation in a dose dependent manner (p < 0.001).

Figure 3.6. ATRA enhances GN11 neuron migration and/or proliferation. Number of neurons in area scraped at different concentrations (vehicle, 0.25, 0.5 or 1 µM) of RA (n=3), over the course of 3 days; p < 0.001; 2-way repeated measures ANOVA.

DISCUSSION

Morphogen gradients during development are crucial organizing centers for embryonic patterning and proper gene expression. In the present study, we show that RA is secreted in the embryonic OP and plays a role in GnRH mRNA transcription as well as cell migration.

However, we reject our initial hypothesis that RA acts as a negative repressor of FGF8. Despite evidence that RA can directly repress FGF8 in the embryonic trunk, we show that RA does not act directly on FGF8 in the OP. However, our results indicate that repression of RA synthesis by citral can affect Fgf8 mRNA transcription, likely through an upstream mechanism. Interestingly, we also found that in the presence of 1 µM ATRA, GnRH mRNA was altered in an opposing manner. In OP cultures treated with 1 µM ATRA, GnRH mRNA decreased, however in the presence of citral, GnRH mRNA also decreased. These results suggest a dynamic pathway of upstream factors controlling GnRH transcription. In addition to these studies, we also sought to understand the role of RA on GnRH neuron function. Accordingly, we challenged the migratory ability of GnRH neurons in the presence of ATRA. We showed that ATRA could increase the number of neurons migrating and/or proliferating into the center of the plate in a dose response manner. Overall, these results resolve and highlight the dynamics of FGF and RA signaling in the OP, and their effect on GnRH neuron migration.

While RALDH2 mRNA expression in the OP significantly increased over embryonic age, we did not detect a change in Fgf8 mRNA in the presence of ATRA in GN11s or in the OP.

However, we found that citral could increase Fgf8 mRNA in the OP. Citral is a competitive inhibitor of RALDHs, which has been effectively used to block RA synthesis in multiple systems

(Kikonyogo et al., 1999; Mey et al., 1997; Thiede et al., 2014). While the addition of ATRA did not directly affect Fgf8 mRNA transcription, the absence of RA synthesis could increase Fgf8

85 mRNA transcription, likely through an indirect mechanism. It is also likely that in the absence of

RA synthesis, Fgf8 mRNA transcription remains high to maintain neurons in an undifferentiated state. This premise would also explain why GnRH mRNA transcription was significantly higher in citral treated OP cultures. Overall, we conclude that RALDH activity is required for Fgf8 transcription, and that in early OP development, ATRA suppresses GnRH mRNA production.

Our ChIP results indicate that endogenous RARs do not bind the FGF8 RARE site at

E9.5 or E13.5, which indicate that RA does not directly affect FGF8. These results follow our

Fgf8 expression experiments in GN11 neurons and in the OP. We also tested the abundance of repressive histone modifications on the RARE site, because previous experiments have demonstrated the ability of RARs to recruit H3K27me3 (Kumar & Duester, 2014). Interestingly, we found that at E13.5, H3K27me3 was associated with the RARE site, despite the absence of

RAR. While no other RARE sites are located on the Fgf8 promoter, there are many putative

RXR sites which can bind RXR-RAR heterodimers, however, for these studies, alternative sites were not tested.

The effect of ATRA inducing GnRH neuron migration in cell scraping assays were particularly interesting due to the confounding effects from OP culture treatment with ATRA and citral. Previous experiments in the embryonic OP with RA-coated beads demonstrated that RA can repress early GnRH-1 positive neurons, however at later developmental time periods, this effect was lost (Sabado et al., 2012). Likewise, it is well known FGF8 signaling is required for early specification of GnRH neurons however, later FGF inhibitor treatments have no significant effect on GnRH-1 positive neurons (Sabado et al., 2012). Our results parallel these findings and raise the possibility that while FGF8 signaling is required for GnRH neuron fate-specification,

RA signaling acts later in GnRH neuron development. The function of RA during OP

86 development may instead be explained by a neuronal differentiation pathway or a migratory pathway. Interestingly, several studies have demonstrated the ability of RA to induce neuronal migration in various types of neurons including GABAergic interneurons, Schwann cells, and mesenchymal stem cells through activation of nuclear RARs or downstream inhibition of caspase activity (Crandall et al., 2011; Latasa et al., 2016; Pourjafar et al., 2017). Indeed, we found that

RA could increase the migration and/or proliferation of GnRH neurons into the scraped area in a dose-dependent manner. These results align with our discovery that RALDH2 mRNA expression is upregulated 400-fold during GnRH neuron onset and may be important for local synthesis of

RA during migration.

In conclusion, we present evidence for an indirect role of RA in the developing mouse

OP. The precise molecular mechanisms of RA on target genes during GnRH neuron development have yet to be fully elucidated. We for the first time, show that RA can control migration and/or proliferation of developing GnRH neurons in vitro. Further perspectives of the functional significance as it pertains to disorders of the congenital endocrine system require further analysis.

REFERENCES

Ang, H. L., Deltour, L., Žgombić‐Knight, M., Wagner, M. A., & Duester, G. (1996). Expression

patterns of class I and class IV alcohol dehydrogenase genes in developing epithelia

suggest a role for alcohol dehydrogenase in local retinoic acid synthesis. Alcoholism:

Clinical and Experimental Research, 20(6), 1050-1064.

Cho, S., Cho, H., Geum, D., & Kim, K. (1998). Retinoic acid regulates gonadotropin-releasing

hormone (GnRH) release and gene expression in the rat hypothalamic fragments and

GT1-1 neuronal cells in vitro. Brain Res Mol Brain Res, 54(1), 74-84. doi:

10.1016/s0169-328x(97)00325-2

Cho, S., Chung, J., Han, J., Ju Lee, B., Han Kim, D., Rhee, K., & Kim, K. (2001). 9-cis-Retinoic

acid represses transcription of the gonadotropin-releasing hormone (GnRH) gene via

proximal promoter region that is distinct from all-trans-retinoic acid response element.

Brain Res Mol Brain Res, 87(2), 214-222. doi: 10.1016/s0169-328x(01)00020-1

Crandall, J. E., Goodman, T., McCarthy, D. M., Duester, G., Bhide, P. G., Drager, U. C., &

McCaffery, P. (2011). Retinoic acid influences neuronal migration from the ganglionic

eminence to the cerebral cortex. J Neurochem, 119(4), 723-735. doi: 10.1111/j.1471-

4159.2011.07471.x

Das, B. C., Thapa, P., Karki, R., Das, S., Mahapatra, S., Liu, T. C., . . . Evans, T. (2014).

Retinoic acid signaling pathways in development and diseases. Bioorg Med Chem, 22(2),

673-683. doi: 10.1016/j.bmc.2013.11.025

Dubey, A., Rose, R. E., Jones, D. R., & Saint-Jeannet, J. P. (2018). Generating retinoic acid

gradients by local degradation during craniofacial development: One cell's cue is another

cell's poison. Genesis, 56(2). doi: 10.1002/dvg.23091

Evans, R. M., & Mangelsdorf, D. J. (2014). Nuclear Receptors, RXR, and the Big Bang. Cell,

157(1), 255-266. doi: 10.1016/j.cell.2014.03.012

Giguere, V., Ong, E. S., Segui, P., & Evans, R. M. (1987). Identification of a receptor for the

morphogen retinoic acid. Nature, 330(6149), 624-629. doi: 10.1038/330624a0

Huang, H. S. F., Dyrenfurth, I., Gunsalus, G. L., & Hembree, W. C. (1985). Effect of Vitamin A

Deficiency upon Gonadotropin Response to Gonadotropin-releasing Hormone1. Biology

of Reproduction, 33(5), 1176-1187. doi: 10.1095/biolreprod33.5.1176

Kanai, M., Raz, A., & Goodman, D. S. (1968). Retinol-binding protein: the transport protein for

vitamin A in human plasma. J Clin Invest, 47(9), 2025-2044. doi: 10.1172/jci105889

Kikonyogo, A., Abriola, D. P., Dryjanski, M., & Pietruszko, R. (1999). Mechanism of inhibition

of aldehyde dehydrogenase by citral, a retinoid antagonist. Eur J Biochem, 262(3), 704-

712. doi: 10.1046/j.1432-1327.1999.00415.x

Kumar, S., & Duester, G. (2014). Retinoic acid controls body axis extension by directly

repressing Fgf8 transcription. Development, 141(15), 2972-2977. doi:

10.1242/dev.112367

Latasa, M. J., Jimenez-Lara, A. M., & Cosgaya, J. M. (2016). Retinoic acid regulates Schwann

cell migration via NEDD9 induction by transcriptional and post-translational

mechanisms. Biochim Biophys Acta, 1863(7 Pt A), 1510-1518. doi:

10.1016/j.bbamcr.2016.04.009

Livak, K. J., & Schmittgen, T. D. (2001). Analysis of relative gene expression data using real-

time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods, 25(4), 402-408.

doi: 10.1006/meth.2001.1262

Mellon, P. L., Windle, J. J., Goldsmith, P. C., Padula, C. A., Roberts, J. L., & Weiner, R. I.

(1990). Immortalization of hypothalamic GnRH neurons by genetically targeted

tumorigenesis. Neuron, 5(1), 1-10.

Mey, J., McCaffery, P., & Dräger, U. C. (1997). Retinoic Acid Synthesis in the Developing

Chick Retina. The Journal of Neuroscience, 17(19), 7441-7449. doi:

10.1523/jneurosci.17-19-07441.1997

Niederreither, K., Fraulob, V., Garnier, J. M., Chambon, P., & Dolle, P. (2002). Differential

expression of retinoic acid-synthesizing (RALDH) enzymes during fetal development

and organ differentiation in the mouse. Mech Dev, 110(1-2), 165-171. doi:

10.1016/s0925-4773(01)00561-5

Niederreither, K., McCaffery, P., Drager, U. C., Chambon, P., & Dolle, P. (1997). Restricted

expression and retinoic acid-induced downregulation of the retinaldehyde dehydrogenase

type 2 (RALDH-2) gene during mouse development. Mech Dev, 62(1), 67-78. doi:

10.1016/s0925-4773(96)00653-3

Petkovich, M., Brand, N. J., Krust, A., & Chambon, P. (1987). A human retinoic acid receptor

which belongs to the family of nuclear receptors. Nature, 330(6147), 444-450. doi:

10.1038/330444a0

Pourjafar, M., Saidijam, M., Etemadi, K., & Najafi, R. (2017). All-trans retinoic acid enhances in

vitro mesenchymal stem cells migration by targeting matrix metalloproteinases 2 and 9.

Biotechnol Lett, 39(8), 1263-1268. doi: 10.1007/s10529-017-2350-1

Sabado, V., Barraud, P., Baker, C. V., & Streit, A. (2012). Specification of GnRH-1 neurons by

antagonistic FGF and retinoic acid signaling. Dev Biol, 362(2), 254-262. doi:

10.1016/j.ydbio.2011.12.016

Sirbu, I. O., & Duester, G. (2006). Retinoic-acid signalling in node ectoderm and posterior

neural plate directs left-right patterning of somitic mesoderm. Nat Cell Biol, 8(3), 271-

277. doi: 10.1038/ncb1374

Thiede, B. R., Mann, Z. F., Chang, W., Ku, Y. C., Son, Y. K., Lovett, M., . . . Corwin, J. T.

(2014). Retinoic acid signalling regulates the development of tonotopically patterned hair

cells in the chicken cochlea. Nat Commun, 5, 3840. doi: 10.1038/ncomms4840

Yoshida, A., Rzhetsky, A., Hsu, L. C., & Chang, C. (1998). Human aldehyde dehydrogenase

gene family. Eur J Biochem, 251(3), 549-557. doi: 10.1046/j.1432-1327.1998.2510549.x

Zhao, D., McCaffery, P., Ivins, K. J., Neve, R. L., Hogan, P., Chin, W. W., & Drager, U. C.

(1996). Molecular identification of a major retinoic-acid-synthesizing enzyme, a

retinaldehyde-specific dehydrogenase. Eur J Biochem, 240(1), 15-22. doi:

10.1111/j.1432-1033.1996.0015h.x

CHAPTER IV

TET1 REGULATES FIBROBLAST GROWTH FACTOR 8 TRANSCRIPTION IN GONADOTROPIN RELEASING HORMONE NEURONS. This work has been published previously as “ TET1 regulates fibroblast growth factor 8 transcription in gonadotropin releasing hormone neurons.” Linscott et. al, PLOS ONE, 2019, 14(7): e0220530.

INTRODUCTION

Fibroblast growth factors (FGFs) are well-known signaling proteins that are crucial for neuronal fate specification, progenitor cell proliferation, and cell survival (Chung et al.,

2010; W. C. J. Chung et al., 2008; Dode et al., 2003; J. C. Gill & P.-S. Tsai, 2006;

Ornitz, 2000; Stevenson et al., 2013; P.-S. Tsai et al., 2011). In the developing brain,

FGF8 is required for proper formation of the midbrain-hindbrain, telencephalon, midline structures, cerebellum, and the olfactory placode (OP) (Chi et al., 2003; McCabe et al.,

2011; Stewart et al., 2016; Theil et al., 2008; Xu et al., 2000). Indeed, inactivation of

FGF8 function results in malformation of various brain regions (Chung & Tsai, 2010;

Creuzet et al., 2004; Lim et al., 2015; Sunmonu et al., 2011; Toyoda et al., 2010; Walshe

& Mason, 2003). As such, location and dosage of Fgf8 mRNA expression is critical for initiating developmental cellular responses, as they elicit downstream signaling factors, which in turn, establish patterning and orientation of brain regions (Alexandre et al.,

2006; Fukuchi-Shimogori & Grove, 2001; Storm et al., 2006; Suzuki-Hirano &

Shimogori, 2009; Toyoda et al., 2010; Xu et al., 2000).

Our previous studies showed that FGF8 signaling is required for gonadotropin-releasing hormone (GnRH) neuron ontogenesis in the OP (Chung et al., 2010; Chung & Tsai, 2010; W. C.

J. Chung et al., 2008; P.-S. Tsai et al., 2011). However, it is still not understood what drives Fgf8 transcription in the developing OP. Several studies showed that multiple downstream, conserved

DNA sequences could recapitulate Fgf8 expression patterns in both zebrafish and mouse embryos (Beermann et al., 2006; Inoue et al., 2006; Marinic et al., 2013). These studies suggested chromosomal conformation is dynamically regulated in the context of tissue type, and that both activating and repressive cis-regulatory elements can drive Fgf8 expression. However, few studies have addressed the chromatin state of the Fgf8 locus, which is likely to play a role in cis-regulatory enhancer-promoter interactions, or alternatively, how the epigenomic state may control the temporal regulation of Fgf8 transcription.

Our recently published results studied Fgf8 gene transcription during emergence of GnRH neurons in the embryonic day (E) E9.5-E13.5 OP, a heterogeneous cell population of neurons and epithelial cells, that contributes to GnRH neuron proliferation and differentiation (M. L.

Linscott & W. C. Chung, 2016; Schwanzel-Fukuda & Pfaff, 1989; Wray, Grant, et al., 1989)

(Fig. 4.1.). In these studies, we found that Fgf8 transcription may be under the direct control of

DNA methylation (Chung & Auger, 2013). Interestingly, we found that the Fgf8 promoter region and gene body harbors three major CpG islands upstream (CpG 1, 2, 3) and one downstream of the transcription start site (TSS; CpG 4) of the Fgf8 gene. Moreover, our experimental data showed that the DNA methyltransferase (DNMT) inhibitor, 5-azacitidine (AZA), induced Fgf8 mRNA expression in GT1-7 GnRH neurons (M. L. Linscott & W. C. Chung, 2016).

Collectively, the results from these initial studies led to the premise that the upregulation of Fgf8 transcription in the embryonic mouse OP may be a DNA methylation-dependent process.

Figure 4.1. A) Schematic of transient Fgf8 transcription during GnRH neuronal emergence in the embryonic mouse OP. B) Fgf8 and GnRH mRNA expression at E10.5 (n=2) or at E10.5+3 days in-vitro (DIV) (n=3) in the mouse OP. p < 0.0005, p < 0.01; Student’s t-test C) MethPrime

CpG prediction of CpG islands on mouse Fgf8 promoter with relative locations of CpG1-4 primers (Li & Dahiya, 2002). * indicates p <0.05.

To test our hypothesis, we focused on two mechanisms that contribute to methylation:

CpG dinucleotide methylation and histone modifications. The ability of DNA and histone methylating proteins to modify chromatin architecture is a key developmental regulatory component, which maintain genes in an active or inactive state, and control cell fate decisions. In general, DNA methylation changes are catalyzed by DNMTs, which convert cytosines to 5- methylcytosine (5mC), and methylated DNA can be demethylated by ten-eleven translocation methylcytosine dioxygenases (TETs), which convert 5mC to 5-hydroxymethylcytosine (5hmC)

(Wu & Zhang, 2017). In neurons, the transition from progenitor to a differentiated neuron is associated with high levels of 5hmC (M. A. Hahn et al., 2013; Santiago et al., 2014; Sun et al.,

2014). At the genomic level, it is well known that 5hmC associates with genes important for neuronal function and correspond with gene transcription (M. A. Hahn et al., 2013; Lister et al.,

As previously indicated, the methylation pattern of histones is of equal importance in respect to Fgf8 transcription. Specifically, two modifications, H3K27me3, a repressive histone mark, and H3K4me3, an activating histone mark, form bivalent domains which poise neurogenic genes for activation, and are thought to respond to developmental cues. Chromatin-associated proteins, such as the polycomb repressive complex 2 (PRC2), which contribute to H3K27 trimethylation, are also important for developmental gene regulation, as they have been shown to regulate proliferation, neurogenesis, WNT signaling, cell cycle exit, and compaction of chromatin (Zemke et al., 2015). Interestingly, mice mutant for Fgf8 or PRC2 genes (i.e., histone-lysine N-methyltransferase Ezh2) share a common phenotype in which embryos fail to

95 undergo gastrulation and have reported proliferation defects (Guo & Li, 2007; O'Carroll et al.,

2001; Sun et al., 1999). Additionally, Fgf8 is suppressed in the trunk tissue of developing embryos (Kumar & Duester, 2014) by a mechanism that recruits PRC2 proteins. These studies suggest a mechanism for Fgf8 transcription via local histone modifications that may affect the accessibility of the chromatin. However, little is known about whether PRC2 proteins also regulate OP Fgf8 transcription during GnRH neuron emergence, and whether they contribute to

GnRH neurobiology.

Here we show that Fgf8 expression in the developing mouse OP is under the control of epigenetic switches involving both DNA and histone-modifying proteins. Specifically, we studied epigenetic control of Fgf8 transcription in the E9.5 - E13.5 mouse OP, which is known to have high Fgf8 transcriptional activity. First, we examined DNMT and TET mRNA expression as well as protein-DNA binding profiles, and whether inhibition of DNMTs upregulated Fgf8 mRNA in embryonic OP explants. Second, we investigated 5hmC conversion via TET1 on CpG islands along the Fgf8 promoter during OP cell differentiation. Specifically, we focused on

TET1, since unlike the other TETs, it is associated with gene promoters (Huang et al., 2014).

Third, we demonstrated that Fgf8 histone modifications, along with other DNA modifiers, are essential for the precise timing of Fgf8 mRNA production during mid-gestational mouse development. Together, we show that the state of Fgf8 chromatin contributes to temporally dependent regulation of transcriptional activity.

MATERIALS AND METHODS

Timed-breeding of mice and nasal explant cultures

Adult wildtype 129P2/OlaHsd*CD-1 male x female mice were timed-bred in the late afternoon in our animal facility (12L:12D cycle) with access to food and water ad libitum. All procedures were approved by the Institutional Animal Care and Use Committee at Kent State University. In the morning, females with a sperm plug were denoted as embryonic day (E) 0.5. Adult pregnant female 129P2/OlaHsd*CD-1 mice were sacrificed at E9.5, E10.5 or E13.5. The uterine horns were quickly removed from the mice and kept in sterile ice-cold phosphate-buffered saline

(Sigma-Aldrich, P3813). Following, the nasal region containing the OPs was surgically isolated using a dissection microscope, placed on 0.65 µm Durapore membrane filters (Millipore,

DVPP04700), transferred to cell tissue culture inserts (Corning, 353095), and grown using the liquid-air interphase method with in phenol-red free Dulbecco's modified Eagle's medium

(DMEM)/F12/glutamax (Thermofisher Scientific, 10565018) supplemented with B27

(Thermofisher Scientific, 17504044) and 1% pen/strep/myc (Sigma-Aldrich, A5955) media. For expression assays of TETs and DNMTs, Chromatin immunoprecipitation, and MeDIP experiments, E9.5, E10.5, or E13.5 nasal explant tissues were collected, immediately flash frozen and kept at -80C.

Cell Culture

Immortalized mouse GnRH GT1-7 and GN11 neurons (generously donated by Dr. Pamela

Mellon, University of San Diego, CA; generously donated by Dr. Sally Radovick, Robert Wood

Johnson Medical School, NJ) were grown in phenol-red free DMEM containing 4.5 g/L pyruvate and 548 mg/L-glutamine, 10% fetal bovine serum (ATCC, 30-2020), 1% pen/strep (Sigma-

Aldrich, A5955) (Mellon et al., 1990). Cells were kept in a humidified incubator at 37°C with

5% CO2.

Pharmacological treatments

OP explants were treated for 72 h (i.e., 3 days in vitro (DIV)) in the presence of vehicle (0.005%

DMSO) or 1 μM 5-azacitidine (AZA) (Tocris Biosciences, 3842). Our AZA dose and length of treatment was based on previous studies showing that these conditions were able to induce gene expression (Y. J. Xin et al., 2015; Zhou & Hu, 2015). Furthermore, AZA was able to induce

Fgf8 mRNA expression in GT1-7 neurons from our previous studies (M. L. Linscott & W. C.

Chung, 2016). Total cellular RNA extraction and cDNA synthesis were performed as described below.

RT-qPCR

Strand cDNA Synthesis Kit (New England Biolabs, E6560L). ProtoScript® II First Strand cDNA Synthesis Kit was used to reverse transcribe 0.5 µg of total RNA. RT-qPCR was performed in triplicate with gene-specific, intron-spanning primers (Table 4.1) using a

Mastercycler EP Realplex2 (Eppendorf, EPPE6300000.604) with SYBR Green PCR Master Mix

(Roche, 04707516001). Relative mRNA expression levels were calculated using the ΔΔ-2CT method (Livak & Schmittgen, 2001). Hypoxanthine phosphoribosyltransferase 1 (Hprt-1) was used as a housekeeping gene.

Table 4.1. Forward and reverse primer sequences for detection of genomic DNA and mRNA transcripts used in this study.

Target Forward (5'-3') Reverse (5'-3') bp

DNA

Fgf8 15

CpG 1 AACTGCTCGTGGTCGTACAG GTGCCCCCAACTAACTCCTC 2

Fgf8 15

CpG 2 GGTGGACGTCGAGCACAG AAGGGCTATCCCGAAAAGGTG 6

Fgf8 14

CpG 3 ACATTAGGCGACCCAGAGAC CGGGATCGTCCAGGGATTG 4

Fgf8 27

CpG 4 GGTACAAGGGCAATGGGGAC CACCTTACCGAAGGGGTCTC 5

Fgf8- non- GTCAGTCTGCGAATATAGCTCA CACAGTACCAACAAGTGTCACA 31 specific G G 4

Fgf8 24

3'UTR CCCAACTACCTGCAGAGCAA TTGAGGAACTCGAAGCGCAG 2 mRNA

Tet1 ACAAAAAGCGTACCTGCACC CCGGTTTTCACGTCACTTCC 4

Tet2 AGGGACCAGAACCAGGCT TTGAATGAATCCAGCAGCACC 1

Tet3 CCTCGGCGGGGATAATGG ACGAGCATTTATTTCCACCTCG 78

DNMT3 23 a GAGCCGCCTGAAGCCC TTTCGATCATCCTCCCGCTC 0

DNMT3 29 b ATCCATAGTGCCTTGGGACC CTCCTGTCATGTCCTGCGT 4

DNMT1 GTACATGCTGCTTCCGCTTG CAAGTCTTTGAGCCGCCTG 7

Fgf8 AGAAGACGGAGACCCCTTCG TGAATACGCAGTCCTTGCCTT 8

GnRH GGCATTCTACTGCTGACTGTGT CTACATCTTCTTCTGCCTGGCT 2

Fgfr1 ATGGTTGACCGTTCTGGAAG TGGCTATGGAAGTCGCTCTT 1

CTCATGGACTGATTATGGACAG GCAGGTCAGCAAAGAACTTAT 12

Hprt GAC AGCC 3

100

Methylated immunoprecipitation (MeDIP)

Olfactory placodes from 4-6 E9.5, E10.5, and E13.5 embryos were treated with RNaseA, lysed overnight and isolated with phenol chloroform. 5 µg of DNA was sonicated to 500-200 bp. DNA was incubated at 95°C for 10 minutes and quickly placed on ice. Dynabeads (Invitrogen,

11203D) were incubated for 2 hours with salmon sperm and washed with PBS-T. DNA (800 ng) was immunoprecipitated with 1.25 µg of either rabbit anti-5mC (Active Motif, 61255), rabbit anti-5hmC (Active Motif, 39791), or normal rabbit IgG (Millipore, 12-370), and 8 ng (1%) of the precipitated DNA was saved as input control. Following incubation, DNA was incubated for 4 hours at 4°C on a 3D rotator with the respective antibodies. Beads were pelleted using a magnet and washed in PBS and TE buffer on a 3D rotator for 10 minutes at 4°C. Antibodies were digested with PK at 60°C for 2 hours followed by a 10 minute 95°C PK inactivation step.

Samples were subsequently purified using phenol-chloroform and glycogen. Each sample was ran on an Eppendorf RT-qPCR program with primers flanking 4 CpG islands on the Fgf8 gene.

Relative enrichment was calculated using the percent input method, where each immunoprecipitation is adjusted to input loading controls, and compared to negative control IgG for significant 5mC or 5hmC enrichment. Non-specific primers flanking upstream of the of Fgf8 were used as negative control in the DNMT3b ChIP.

Chromatin Immunoprecipitation (ChIP)

ChIP was used to examine histone modifications (H3K4me3 and H3K27me3) along with TET1,

DNMT3a, DNMT3b and DNMT1, and EZH2 interaction along the promoter region of Fgf8.

Embryos were dissected at E9.5, E10.5 or E13.5, and pooled in microcentrifuge tubes containing

4-6 OPs per sample. Chromatin was harvested using the EZ-Magna-ChIP kit (Millipore, 17-

101

10086) according to manufacturer's instructions. Briefly, cells were cross-linked with 1% formalin for 10 min and lysed. The protein cross-linked genomic DNA was fragmented to 200–

600 base pairs using sonication. All chromatin samples were verified for correct shearing density on an agarose gel before continuing. Magnetic Protein A/G beads were blocked with salmon sperm for 2 hours and washed in PBS. Following, the fragments were immunoprecipitated using

1.25 μg of either rabbit polyclonal antibody against DNMT3b, DNMT1, DNMT3a (Abcam, ab2851, ab13537, ab2850 ), EZH2, TET1, H3K4me3, and H3K27me3 (Active Motif, 39901,

61443, 39915, 39155) or control IgG (Millipore, 12-370) for 3 hours at 4°C on a 3D rotator.

Magnetic beads were pulled out of solution using a magnet, and bound fragments were washed four times for ten minutes. Proteinase K (10 mg/ml) was used to reverse crosslinking, and DNA was isolated using phenol chloroform. The relative amount of protein occupancy on the identified sites on the promoter of Fgf8 was measured using a Mastercycler EP

Realplex2 (Eppendorf, EPPE6300000.604) with SYBR Green PCR Master Mix (Roche,

04707516001). For this purpose, four primer sets were designed to flank the CpG islands on the promoter region and an intragenic site of Fgf8. ChIP signal was measured using the percent input method as described above. Non-specific primers flanking the 3’UTR of Fgf8, a CpG poor region, were also used in the MeDIP (S4.11. Fig).

Tet1 siRNA treatment siRNA experiments were conducted using Accell SMARTpool siRNA, which targets 4 portions of the Tet1 mRNA transcript (CUAUUUGUCUAUUAUGUGUG,

UCGUUGGGUCUAAAGGCUU, UCGCUAAACUAACUAUAAAUGUAU,

UUAUAGUUUUAAAUACUUA). A non-targeting siRNA pool was used as a negative control

102

(UGGUUUACAUGUCGACUAA,

UGGUUUACAUGUUUUCUGA,UGGUUUACAUGUUUUCCUA,

UGGUUUACAUGUUGUGUGA). All siRNAs were resuspended in 5x siRNA buffer diluted in

RNase free water. GT1-7 hypothalamic neurons were seeded in 24-well plates in DMEM and

FBS with 1% penicillin/streptomycin. Growth medium was removed and replaced with Accell delivery media (Dharmacon, B-005000-100) with 1 M Tet1 siRNA (Dharmacon, E-06861-00-

0020) or non-targeting control (Dharmacon, D-001910-20-20). After 3 days, cells were collected for RNA as described previously.

5mC and 5hmC Dot Blot

Dot blot experiments were used to determine genome-wide changes in DNA. Total DNA was isolated as previously described. 150ng and 75ng of DNA were diluted in 20x SSC, heated at 95º

C, cooled on ice, and blotted on a nitrocellulose membrane soaked in 10x SSC. The membrane was UV crosslinked for 10 minutes on a Benchtop UV Transilluminator (UVP, M-20V).

Following, the membrane was blocked in 5% dry milk and 2% normal sheep serum for 1 hour, before an overnight incubation (1:5000) with 5hmC or 5mC (Active Motif, 39791, 61255) in blocking solution. The membrane was washed three times in TBS-Tween and incubated in peroxidase anti-rabbit IgG (1:2000) (Vector Laboratories, PI-1000) for one hour, and imaged using Clarity™ Western ECL substrate (BioRad, 1705061). Images were processed and quantified using ImageJ studio.

Statistical analysis

103

Data were analyzed using Student t-tests or one-way analysis of variance (ANOVA) with treatment and/or DIV between subject variables. Holm-Sidak method tests were used for post- hoc analysis. Group numbers (n) are given in the figure legends. Differences were considered significant if p < 0.05.

RESULTS

Identification of CpG islands on the Fgf8 promoter

To identify CpG islands along the length of the Fgf8 promoter, we analyzed 5000 basepairs along the 5’UTR and an intragenic site localized between exons 1 and 2 using MethPrimer (Li &

Dahiya, 2002). Here, we found 4 major CpG islands, with 3 located in the 5’ UTR and 1 located intragenically (Fig. 4.1. C).

Inhibition of DNMT activity increases Fgf8 expression in the embryonic mouse OP

Our previous studies in GT1-7 neurons revealed that inhibition of DNMT activity using AZA, upregulated Fgf8 expression. Here, we found that AZA treatment (1 µM) in E10.5 OP explants cultured 3 DIV increased Fgf8 mRNA expression (p = 0.02; Fig. 4.2. A). Contrary to what was found previously in the GT1-7 cell line, the increase in OP GnRH mRNA expression was not significant (p = 0.1; Fig. 4. 2. A) (M. L. Linscott & W. C. Chung, 2016). Dot blot analysis of overall 5mC levels in GT1-7 cells showed a decrease with increasing AZA concentration (Fig.

4.2. B). Propidium iodine staining in GT1-7 neurons showed no significant changes in cell death when treated with 1 µM AZA 3 DIV (p = 0.5), (S4.12. Fig).

104

Figure 4.2. AZA induced Fgf8 expression in the OP. A) RT-qPCR for Fgf8, GnRH and HDAC1 mRNA in vehicle (n = 6) vs 1 µM AZA-treated E10.5 mouse OP explants (n = 6) for 3 DIV B)

Two representative 5mC dot blots in GT1-7 neurons treated with AZA (Vehicle, 0.01, 0.1, or 1

µM) for 3 days. Note that 5mC is virtually absent at the 1 µM concentration only. * indicates p <

0.05; Student’s t-test.

105

Tet and DNMT expression in the OP

To determine the transcriptional activity of DNA methylation modifiers in the embryonic OP, we analyzed their mRNA levels in E10.5 and E13.5 OPs. We showed that Tet1, Tet2, and Tet3 mRNA was increased in E13.5 OPs compared to E10.5 OPs (p = 0.001, p = 0.0002, p = 0.0006 respectively; Fig. 4.3. A). Dnmt3a mRNA increased 10-fold (p = 0.006), while Dnmt3b mRNA levels significantly decreased (p = 0.03; Fig. 4.3. B). No significant changes were detected for

Dnmt1 or Hdac1 mRNA expression (p = 0.5, p = 0.9). We further found that Tet1, Tet2, and

Tet3 mRNA expression was also higher in developmentally mature GT1-7 compared to migratory, immature GN11 neurons (p = 0.000003, p = 0.0007, p = 0.00002; Fig. 4.3. C).

Moreover, 5hmC DNA dot blot experiments showed that 5hmC levels are significantly higher in

E13.5 compared to E10.5 OPs (p = 0.01; Fig. 4.3. D).

106

Figure 4.3. TET or DNMT expression in the OP. A) Tet 1, 2, 3 or B) Dnmt1, 3a, 3b, and HDAC1 mRNA expression in E10.5 or E13.5 mouse OP (n = 4). C) Tet 1, 2, 3 mRNA expression in

GN11 and GT1-7 GnRH neurons (n = 4). D) 5hmC dot blot quantification of E10.5 (n = 4) versus E13.5 (n = 4) in 75 ng of OP genomic DNA and original dot blot image. * indicates p <

0.05; Student’s t-test.

107

DNMT3b interacts with the Fgf8 locus

Our AZA treatment findings suggested that Fgf8 expression was under the control of repressive

DNA methylation in both the GT1-7 cell line and E10.5 OP. To explore this further, we tested the ability of DNMT proteins to associate with the Fgf8 locus. While DNMT3a and DNMT1 were not bound to the Fg8 locus at E9.5 or E13.5, DNMT3b was significantly enriched at E9.5 on CpG 2 (p = 0.03), CpG 3 (p = 0.03) and CpG 4 (p = 0.03), which are closest to the Fgf8 TSS.

In contrast, DNMT3b was not significantly enriched on CpG 1 (p = 0.06) and a non-specific region (p = 0.8) (Fig. 4.4.).

108

Figure 4.4. DNMT3b binds to the Fgf8 promoter early in OP development. DNMT3b ChIP-RT- qPCR of 6 pooled E9.5 mouse OPs on the Fgf8 promoter (n = 4). * indicates p < 0.05. Student’s t-test.

109

5hmC accumulation on the Fgf8 promoter is driven in a time-dependent fashion

Here, we analyzed 5mC and 5hmC occupation along the Fgf8 promoter region in developing

OPs. Because our previous results indicated rapid changes in the expression levels of epigenetic modifiers (i.e., increased Tet and Dnmt3a and reduced Dnmt3b mRNA levels), we hypothesized that these enzymes may play a prominent role in regulating Fgf8 transcription. We performed

MeDIP using OPs from E9.5, E10.5 and E13.5 embryos. We found that 5hmC was enriched at 2 of the 3 CpG islands found along the promoter region of Fgf8 (Fig. 4.5.). Individual Student’s t- tests comparing 5hmC levels within CpG 1 or CpG 3 found that 5hmC was highly enrichment at all ages (E9.5, E10.5, E13.5) when compared to IgG (p = 0.002, p = 0.03, p = 0.03; p = 0.0001, p

= 0.03, p = 0.04; Fig. 4.5.). These results are in line with E8.5 frontonasal prominence tissue

(i.e., pre-placodal) MeDIP data detecting significant 5mC and 5hmC enrichment on CpG 1 (p =

0.01; p = 0.002), and CpG 3(p = 0.0006; p = 0.0006) (S4.13. Fig). Interestingly, this observation coincided with the embryonic time period where OP Fgf8 mRNA expression is high. One-way

ANOVAs showed a significant decrease in 5hmC with time within CpG 1 (p = 0.002) and CpG 3

(p = 0.0001). In contrast, no significant enrichment of 5hmC was detected on CpG 2 or CpG 4 (p

= 0.1, p = 0.2, p = 0.1; p = 0.2, p = 0.5, p = 0.7), and 5mC was not enriched at any of the 4 CpG sites (p = 0.7, p = 0.1, p = 0.6; p = 0.7, p = 1, p = 0.3; p = 0.2, p = 0.2, p = 0.9; p = 0.4, p = 0.8, p

= 0.5, listed as age groups within primer sets, CpG 1-4, respectively.) (Fig. 4.5.). These results were verified using 2 different antibodies against 5mC and 5hmC.

110

Figure 4.5. 5hmC accumulation on the Fgf8 promoter is driven in a time-dependent fashion.

MeDIP RT-qPCR in 3-5 pooled mouse OPs at E9.5 (n = 4), E10.5 (n = 4), E13.5 (n = 4) along the Fgf8 promoter. CpG islands are indicated in numbers 1 - 4. * indicates p < 0.05 compared to

IgG; Student’s t-test. ** indicates p < 0.05 5hmC enrichment on CpG 1 and 3 between E9.5 and

E10.5 or E13.5; One-way ANOVA followed by Holm-Sidak post hoc.

111

TET1 interacts with CpG 1 and CpG 3

Because we found significant enrichment of 5hmC on CpG 1 and CpG 2, we hypothesized that

TET proteins are responsible for the conversion of 5mC to 5hmC on the Fgf8 promoter.

Specifically, we focused on TET1, since it is associated with gene promoters while TET2 is associated with gene bodies and near actively expressed exons (Huang et al., 2014). Moreover,

TET1 seems to play a role in the neuroendocrine system, as recent studies found that TET1-KO mice have impaired fertility, and TET1 expression responds to luteinizing hormone in gonadotropin cell lines (Dawlaty et al., 2011; Yosefzon et al., 2017). Therefore, we performed

ChIP to determine if TET1 co-localizes with 5hmC rich regions on the Fgf8 promoter. Our results showed that at E9.5, TET1 localized to CpG 1 and 3, previously identified to be 5hmC enriched (p = 0.01, p = 0.004), and was absent on CpG 2 (p = 0.7) (Fig. 4.6. A). At E13.5, TET1 was significantly enriched only on CpG 1 (p = 0.03) and 3 (p = 0.04) (Fig. 4.6. B).

112

Figure 4.6. Epigenetic switch on the Fgf8 promoter in the E9.5 and E13.5 OP. A) At E9.5 (n =

4), TET1 was enriched at CpG 1 and 3. B) At E13.5 (n = 4), TET1 was enriched at CpG 1 and 3, while EZH2 was enriched at all 3 CpG sites. * indicates p < 0.05; Student’s t-test.

113

Histone modifications regulate timing of Fgf8 transcription

Because TETs have been shown to functionally associate with PRC2 members and bivalent promoters, of which Fgf8 is known to be part of (Neri et al., 2013; Verma et al., 2018; Zhang &

Reinberg, 2001), we hypothesized that PRC2-dependent histone modifications also contribute to

Fgf8 mRNA transcription. Indeed, our results demonstrate that at E9.5, Fgf8 harbors both

H3K4me3 and H3K27me3 near the TSS (p = 0.004, p = 0.004; Fig. 4.7. A), while at E13.5 only

H3K27me3 is present (p = 0.002). H3K4me3 was not detected at E13.5 (p = 0.6); Fig. 4.7. B).

We also tested whether PRC2 protein, EZH2, was recruited to the Fgf8 promoter, and found that

EZH2, which is responsible of H3K27 trimethylation, was enriched at CpG 1, 2, and 3 on the

Fgf8 promoter (p = 0.02, p = 0.008, p = 0.009; Fig. 4.6. B), which was not the case for E9.5 OPs

(p = 0.1, p = 0.7, p = 0.1; Fig. 4.6. A).

114

Figure 4.7. Fgf8 histone modifications during GnRH neuron ontogenesis. A) ChIP for

H3K4me3/H3K27me3 of 3 - 5 pooled E9.5 (n = 4) and B) 3 - 5 pooled E13.5 OP (n = 4) on CpG

3. E9.5 OPs are enriched for H3K4me3 and H3K27me3, whereas only H3K27me3 was detected in E13.5. * indicates p < 0.05; Student’s t-test.

115

TET1 regulates Fgf8 and Fgfr1 transcription

To determine the regulatory potential of TET1 on Fgf8 mRNA expression, we used Tet1 siRNA to reduce Tet1 expression in GT1-7 neurons, which we previously showed to have higher levels of Tet1 than GN11 neurons (Fig. 4.3. C). Accell Tet1 siRNA experiments demonstrated a ~80% reduction in Tet1 mRNA (p = 0.0002; Fig. 4.8. A) transcript compared to non-targeting control siRNA, and did not affect Tet2 or Tet3 mRNA (p = 0.05, p = 0.2). Accell Tet1 siRNA-treated

GT1-7 neurons showed a significant reduction in Fgf8 and Fgfr1 mRNA (p = 0.04, p = 0.02; Fig.

4.8. B). While GnRH mRNA trended down, it was not significant (p = 0.1).

116

Figure 4.8. Tet1 siRNA knockdown in GT1-7 neurons. A) Tet1 siRNA (n = 4) did not affect other TET or DNMT mRNA expression, and reduced Tet1 mRNA expression compared to non- targeting controls (n = 4). B) Tet1 knockdown reduced Fgf8, Fgfr1, and Fgf2 mRNA (n = 4). * indicates p < 0.05; Student’s t-test.

117

DISCUSSION

Timing of gene expression during development is critical for proper onset of neuronal systems and embryonic patterning. In the present study, we show that Fgf8 transcription is temporally regulated via TET1 during development of the GnRH system in the embryonic mouse OP. We also found that TET1 continued to interact with specific CpG-rich regions on the Fgf8 promoter after emergence of GnRH neurons. Interestingly, we found that these TET1-interacting Fgf8 promoter regions also recruited EZH2, which likely lead to the observed increased H3K27 trimethylation. Taken together, these sequential epigenetic events suggest that TET1 maintains the Fgf8 promoter in a hypomethylated state, while subsequent recruitment of members of the

PRC2 complex promote the repressive actions of H3K27me3 on Fgf8 transcription. These results underscore the importance of epigenetic-dependent timing of Fgf8 expression during GnRH neuron emergence, and that disruptions in the epigenome not only disrupt Fgf8 signaling but may also play a critical role that results in KS pathogenesis.

CpG dinucleotide methylation is largely determined by the activity of DNMTs. While

DNMT1 was not significantly upregulated during OP development, DNMT3a was upregulated ten-fold, while DNMT3b was downregulated in the E13.5 OPs as compared to E10.5 OPs.

Additionally, we found enrichment of 5mC on the Fgf8 promoter at CpG 1 and CpG 3 in E8.5 frontonasal prominences (S4.13. Fig), indicating a role for DNMTs. Previous studies have shown

DNMT3a is detected primarily in post-mitotic olfactory receptor neurons, while DNMT3b expression was restricted to proliferating progenitor neurons, underlining the possibility that

GnRH progenitors or their neural progenitors may, in part, rely on epigenetic machinery as a mechanism for neuronal differentiation (MacDonald et al., 2005). In support, we found that

DNMT3b was bound near the TSS and at an intragenic site of the Fgf8 promoter at E10.5, but

118 devoid of enrichment at E13.5. When E10.5 OP explants were treated with AZA, Fgf8 transcription increased, indicating at E10.5, Fgf8 transcription is under the repressive effects of

DNMTs, presumably by DNMT3b at the gene body. This assumption is based on our observations that only DNMT3b was found to interact with CpG 3 and CpG 4, which are proximally located near the Fgf8’s TSS and 5’ exon/gene body. Moreover, localization of

DNMT3b at the intragenic CpG 4 region on the actively transcribed Fgf8 locus may further suggest CpG 4 is involved with mRNA processing, such as elongation, splicing, or recruitment of other DNA binding proteins (Gatto et al., 2017; Lee et al., 2015; Maunakea et al., 2013; Schor et al., 2010).

Timing of gene expression during embryonic brain development is imperative during fate specification of neuronal populations, of which Fgf8 is critically important. During development of the OP, we found higher expression in Tet1-3 mRNA at E13.5. Furthermore, dot blot analysis showed higher levels of genome-wide hydroxymethylation at E13.5. In the OP, we found a clear relationship between the abundance of Fgf8 mRNA transcripts and 5hmC levels on CpG 1 and 3 of the Fgf8 promoter. These two 5hmC-rich regions were also co-occupied by TET1, demonstrating that TET1 is likely responsible for converting 5mC to 5hmC. These results indicate that TET1 discretely controls the timing of Fgf8 expression during development of

GnRH neurons to ensure proper FGF8 signaling. Interestingly, we also found high 5hmC levels on the Fgf8 promoter in midbrain-hindbrain tissue at E10.5 (S4.14. Fig), which may suggest that

TET-dependent demethylation is a general mechanism for Fgf8 transcription in neuronal populations.

TET-catalyzed demethylation also affects nearby nucleosome compaction by modifying histone methylation status (Juan et al., 2016; Mendonca et al., 2014; Ngo et al., 2016; Teif et al.,

119

2014). Aside from catalyzing the conversion from 5mC to 5hmC, TET1 can also maintain histone bivalency (Gu et al., 2018; Kong et al., 2016; Sui et al., 2012; Verma et al., 2018), which we hypothesized contributes to timing of Fgf8 expression. We therefore measured H3K4 and

H3K27 trimethylation levels in the developing OP. We found that at E9.5, the Fgf8 TSS is associated with H3K4me3 and H3K27me3. In contrast, at E13.5, the Fgf8 TSS exclusively harbors H3K27me3, suggesting that H3K27me3 has a repressive role on Fgf8 transcription.

Based on previous studies, we infer that this histone switch in H3K4 and H3K27 trimethylation promotes nucleosome compaction, and is therefore responsible for the time-dependent downregulation of Fgf8 transcription in the embryonic mouse OP (Wu et al., 2011; Wu &

Zhang, 2011; Zhang & Reinberg, 2001).

Our ChIP studies support that TET1 interaction with the Fgf8 promoter is rather stable and showed continuous interaction with the Fgf8 promoter. We also found that EZH2, which is highly expressed in the E9.5 and also in the E13.5 OP (S4.14. Fig), was enriched on the E13.5, but not on the E9.5 Fgf8 promoter, suggesting that TET1 may help recruit the PRC2 complex to trimethylate H3K27 at these Fgf8 promoter regions. Indeed, previous studies have found that

TET1 protein can directly interact with EZH2 (Cartron et al., 2013; Wu et al., 2011). Taken together these data suggest that TET1 is not only responsible for DNA demethylation, but may also be responsible for subsequent HK27-specific trimethylation through the recruitment of

EZH2, as suggested in earlier studies in mouse embryonic stem cells (Wu et al., 2011; Wu &

Zhang, 2011). It is also possible that TET1 is replaced by EZH2 during periods of high Fgf8 transcription, which will repress Fgf8 transcription by depositing H3K27me3, thereby contributing to the transient nature of Fgf8 expression in the developing mouse OP.

Alternatively, recent biochemical studies indicate a feedback mechanism by which de novo RNA

120 transcripts (including Fgf8 transcripts) contribute to EZH2 recruitment by interacting with the

RNA binding domain of EZH2 (Zhao et al., 2010). In this mechanism, mRNA production is inhibited by recruiting other PRC2 core proteins to gene promoters (Bonasio et al., 2014;

Kaneko et al., 2014; Kaneko et al., 2013). Currently, more studies are needed to pinpoint which of these mechanisms is the most likely pathway that contributes to transcription of Fgf8 in the embryonic mouse OP.

Previous studies in GT1-7 neurons indicated that direct TET function, specifically TET2, is required for GnRH mRNA transcription and affected histone methylation status on the GnRH promoter (Kurian et al., 2016). Here, our Tet1 siRNA experiments demonstrated that TET1 is required for regulating Fgf8 transcription, a molecular event that we previously showed to be required for the emergence of the GnRH neuronal system in mice (W. C. J. Chung et al., 2008;

Falardeau et al., 2008). Therefore, we conclude that TET1 binding to the Fgf8 promoter in OP progenitor cells is not only critical for inducing Fgf8 transcription but may also be required for the emergence of GnRH neurons. Additionally, genome-wide hydroxymethylation of DNA and

RNA, mediated through TET enzyme activity is perhaps of equal importance during development of GnRH neurons. In support, our studies also showed that TET1 controls Fgf2 and

Fgfr1 transcription in GT1-7 neurons. This possibility is not without merit given that earlier studies early showed GnRH neuronal development is a multi-genic process. Overall, our studies provide further evidence that upstream epigenomic regulators are involved in GnRH neuron differentiation and the onset of KS.

In conclusion, the present study provides evidence that Fgf8 is transcriptionally regulated by TET1. We show that TET1 maintains a Fgf8’s hypomethylated state, which decreases with embryonic age. Moreover, TET1 together with EZH2 likely maintains a dynamic bivalent

121 histone state of the chromatin proximal to the Fgf8 locus that contributes to the transient nature of embryonic Fgf8 expression. These and other studies largely point to Fgf8 activation through dynamic chromatin arrangement and specific methylation events (Fig. 4.9.). Furthermore, DNA demethylation and the conversion of 5mC to 5hmC, play an integral role in Fgf8 transcription during development of the GnRH system. Overall, disruptions in the pre-hypothalamic epigenome could have major consequences on the Fgf8 signaling system and GnRH neurodevelopment, resulting in congenital hypogonadotropic hypogonadism disorders, such as

KS.

122

Figure 4.9. Schematic model of age-dependent Fgf8 transcriptional control in the embryonic mouse OP. A) At time (t) = 0, TET1 interaction with the Fgf8 promoter maintains its hypomethylated, and H3K4me3/H3K27me3 bivalent state, thereby inducing Fgf8 transcription

B) In contrast at t = 1, EZH2 recruitment maintained H3K27 trimethylation, while H3K4me3 was lost, which represses Fgf8 transcription. Closed circles = methylated, open circles = demethylated.

123

SUPPORTING INFORMATION

S4.10. Fig. Negative MeDIP qPCR in 3-5 pooled mouse OPs at E9.5 on the 3’UTR Fgf8.

Negative control region in comparison to E9.5 CpG1 site (n=4). * indicates p < 0.05; Student’s t- test.

124

S4.11. Fig. Propidium iodine staining in GT1-7 cells treated with AZA for 3 days in vitro (n =

3); One-way ANOVA.

125

S4.12. Fig. Significant levels of 5mC on the Fgf8 promoter at E8.5. MeDIP qPCR in 8-10 pooled mouse frontonasal prominences at E8.5 along the promoter of Fgf8 CpG 1 and CpG 3 (n=4). * indicates p < 0.05; Student’s t-test.

126

S4.13. Fig. Fgf8 promoter landscape in the E10.5 Midbrain-Hindbrain. MeDIP qPCR in 3-5 pooled mouse midbrain-hindbrains at E10.5 along the promoter of Fgf8 CpG islands indicated in numbers 1-3 (n = 4). * indicates p < 0.05; Student’s t-test.

127

S4.14. Fig. EZH2 is highly expressed in the developing OP. A). EZH2 mRNA expression in the

E9.5 versus E13.5 OP; * indicates p < 0.05; Student’s t-test.

128

REFERENCES

Alexandre, P., Bachy, I., Marcou, M., & Wassef, M. (2006). Positive and negative regulations by

FGF8 contribute to midbrain roof plate developmental plasticity. Development, 133(15),

2905-2913. doi: 10.1242/dev.02460

Beermann, F., Kaloulis, K., Hofmann, D., Murisier, F., Bucher, P., & Trumpp, A. (2006).

Identification of evolutionarily conserved regulatory elements in the mouse Fgf8 locus.

Genesis, 44(1), 1-6. doi: 10.1002/gene.20177

Bonasio, R., Lecona, E., Narendra, V., Voigt, P., Parisi, F., Kluger, Y., & Reinberg, D. (2014).

Interactions with RNA direct the Polycomb group protein SCML2 to chromatin where it

represses target genes. Elife, 3, e02637. doi: 10.7554/eLife.02637

Cartron, P. F., Nadaradjane, A., Lepape, F., Lalier, L., Gardie, B., & Vallette, F. M. (2013).

Identification of TET1 Partners That Control Its DNA-Demethylating Function. Genes

Cancer, 4(5-6), 235-241. doi: 10.1177/1947601913489020

Chi, C. L., Martinez, S., Wurst, W., & Martin, G. R. (2003). The isthmic organizer signal FGF8

is required for cell survival in the prospective midbrain and cerebellum. Development,

130(12), 2633-2644.

Chung, W. C., & Auger, A. P. (2013). Gender differences in neurodevelopment and epigenetics.

Pflugers Arch, 465(5), 573-584. doi: 10.1007/s00424-013-1258-4

Chung, W. C., Matthews, T. A., Tata, B. K., & Tsai, P. S. (2010). Compound deficiencies in

multiple fibroblast growth factor signalling components differentially impact the murine

gonadotrophin-releasing hormone system. J Neuroendocrinol, 22(8), 944-950. doi:

10.1111/j.1365-2826.2010.02024.x

129

Chung, W. C., & Tsai, P. S. (2010). Role of fibroblast growth factor signaling in gonadotropin-

releasing hormone neuronal system development. Front Horm Res, 39, 37-50. doi:

10.1159/000312692

Chung, W. C. J., Moyle, S. S., & Tsai, P.-S. (2008). Fibroblast Growth Factor 8 Signaling

through Fibroblast Growth Factor Receptor 1 Is Required for the Emergence of

Gonadotropin-Releasing Hormone Neurons. Endocrinology, 149(10), 4997-5003. doi:

10.1210/en.2007-1634

Creuzet, S., Schuler, B., Couly, G., & Le Douarin, N. M. (2004). Reciprocal relationships

between Fgf8 and neural crest cells in facial and forebrain development. Proc Natl Acad

Sci U S A, 101(14), 4843-4847. doi: 10.1073/pnas.0400869101

Dawlaty, M. M., Ganz, K., Powell, B. E., Hu, Y. C., Markoulaki, S., Cheng, A. W., . . . Jaenisch,

R. (2011). Tet1 is dispensable for maintaining pluripotency and its loss is compatible

with embryonic and postnatal development. Cell Stem Cell, 9(2), 166-175. doi:

10.1016/j.stem.2011.07.010

Dode, C., Levilliers, J., Dupont, J. M., De Paepe, A., Le Du, N., Soussi-Yanicostas, N., . . .

Hardelin, J. P. (2003). Loss-of-function mutations in FGFR1 cause autosomal dominant

Kallmann syndrome. Nat Genet, 33(4), 463-465. doi: 10.1038/ng1122

Falardeau, J., Chung, W. C., Beenken, A., Raivio, T., Plummer, L., Sidis, Y., . . . Pitteloud, N.

(2008). Decreased FGF8 signaling causes deficiency of gonadotropin-releasing hormone

in humans and mice. J Clin Invest, 118(8), 2822-2831. doi: 10.1172/jci34538

Fukuchi-Shimogori, T., & Grove, E. A. (2001). Neocortex patterning by the secreted signaling

molecule FGF8. Science, 294(5544), 1071-1074. doi: 10.1126/science.1064252

130

Gatto, S., Gagliardi, M., Franzese, M., Leppert, S., Papa, M., Cammisa, M., . . . Matarazzo, M.

R. (2017). ICF-specific DNMT3B dysfunction interferes with intragenic regulation of

mRNA transcription and alternative splicing. Nucleic Acids Res, 45(10), 5739-5756. doi:

10.1093/nar/gkx163

Gill, J. C., & Tsai, P.-S. (2006). Expression of a Dominant Negative FGF Receptor in

Developing GNRH1 Neurons Disrupts Axon Outgrowth and Targeting to the Median

Eminence1. Biology of Reproduction, 74(3), 463-472. doi:

10.1095/biolreprod.105.046904

Gu, T., Lin, X., Cullen, S. M., Luo, M., Jeong, M., Estecio, M., . . . Goodell, M. A. (2018).

DNMT3A and TET1 cooperate to regulate promoter epigenetic landscapes in mouse

embryonic stem cells. Genome Biol, 19(1), 88. doi: 10.1186/s13059-018-1464-7

Guo, Q., & Li, J. Y. (2007). Distinct functions of the major Fgf8 spliceform, Fgf8b, before and

during mouse gastrulation. Development, 134(12), 2251-2260. doi: 10.1242/dev.004929

Hahn, M. A., Qiu, R., Wu, X., Li, A. X., Zhang, H., Wang, J., . . . Lu, Q. (2013). Dynamics of 5-

hydroxymethylcytosine and chromatin marks in Mammalian neurogenesis. Cell Rep,

3(2), 291-300. doi: 10.1016/j.celrep.2013.01.011

Huang, Y., Chavez, L., Chang, X., Wang, X., Pastor, W. A., Kang, J., . . . Rao, A. (2014).

Distinct roles of the methylcytosine oxidases Tet1 and Tet2 in mouse embryonic stem

cells. Proc Natl Acad Sci U S A, 111(4), 1361-1366. doi: 10.1073/pnas.1322921111

Inoue, F., Nagayoshi, S., Ota, S., Islam, M. E., Tonou-Fujimori, N., Odaira, Y., . . . Yamasu, K.

(2006). Genomic organization, alternative splicing, and multiple regulatory regions of the

zebrafish fgf8 gene. Dev Growth Differ, 48(7), 447-462. doi: 10.1111/j.1440-

169X.2006.00882.x

131

Juan, D., Perner, J., Carrillo de Santa Pau, E., Marsili, S., Ochoa, D., Chung, H. R., . . . Valencia,

A. (2016). Epigenomic Co-localization and Co-evolution Reveal a Key Role for 5hmC as

a Communication Hub in the Chromatin Network of ESCs. Cell Rep, 14(5), 1246-1257.

doi: 10.1016/j.celrep.2016.01.008

Kaneko, S., Son, J., Bonasio, R., Shen, S. S., & Reinberg, D. (2014). Nascent RNA interaction

keeps PRC2 activity poised and in check. Genes Dev, 28(18), 1983-1988. doi:

10.1101/gad.247940.114

Kaneko, S., Son, J., Shen, S. S., Reinberg, D., & Bonasio, R. (2013). PRC2 binds active

promoters and contacts nascent RNAs in embryonic stem cells. Nat Struct Mol Biol,

20(11), 1258-1264. doi: 10.1038/nsmb.2700

Kong, L., Tan, L., Lv, R., Shi, Z., Xiong, L., Wu, F., . . . Shi, Y. G. (2016). A primary role of

TET proteins in establishment and maintenance of De Novo bivalency at CpG islands.

Nucleic Acids Res, 44(18), 8682-8692. doi: 10.1093/nar/gkw529

Kumar, S., & Duester, G. (2014). Retinoic acid controls body axis extension by directly

repressing Fgf8 transcription. Development, 141(15), 2972-2977. doi:

10.1242/dev.112367

Kurian, J. R., Louis, S., Keen, K. L., Wolfe, A., Terasawa, E., & Levine, J. E. (2016). The

Methylcytosine Dioxygenase Ten-Eleven Translocase-2 (tet2) Enables Elevated GnRH

Gene Expression and Maintenance of Male Reproductive Function. Endocrinology,

157(9), 3588-3603. doi: 10.1210/en.2016-1087

Lee, S. M., Choi, W. Y., Lee, J., & Kim, Y. J. (2015). The regulatory mechanisms of intragenic

DNA methylation. Epigenomics, 7(4), 527-531. doi: 10.2217/epi.15.38

132

Li, L. C., & Dahiya, R. (2002). MethPrimer: designing primers for methylation PCRs.

Bioinformatics, 18(11), 1427-1431.

Lim, M. S., Lee, S. Y., & Park, C. H. (2015). FGF8 is Essential for Functionality of Induced

Neural Precursor Cell-derived Dopaminergic Neurons. Int J Stem Cells, 8(2), 228-234.

doi: 10.15283/ijsc.2015.8.2.228

Linscott, M. L., & Chung, W. C. (2016). Fibroblast Growth Factor 8 Expression in GT1-7

GnRH-Secreting Neurons Is Androgen-Independent, but Can Be Upregulated by the

Inhibition of DNA Methyltransferases. Front Cell Dev Biol, 4, 34. doi:

10.3389/fcell.2016.00034

Lister, R., Mukamel, E. A., Nery, J. R., Urich, M., Puddifoot, C. A., Johnson, N. D., . . . Ecker, J.

R. (2013). Global Epigenomic Reconfiguration During Mammalian Brain Development.

Science, 341(6146). doi: 10.1126/science.1237905

Livak, K. J., & Schmittgen, T. D. (2001). Analysis of relative gene expression data using real-

time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods, 25(4), 402-408.

doi: 10.1006/meth.2001.1262

MacDonald, J. L., Gin, C. S., & Roskams, A. J. (2005). Stage-specific induction of DNA

methyltransferases in olfactory receptor neuron development. Dev Biol, 288(2), 461-473.

doi: 10.1016/j.ydbio.2005.09.048

Marinic, M., Aktas, T., Ruf, S., & Spitz, F. (2013). An integrated holo-enhancer unit defines

tissue and gene specificity of the Fgf8 regulatory landscape. Dev Cell, 24(5), 530-542.

doi: 10.1016/j.devcel.2013.01.025

133

Maunakea, A. K., Chepelev, I., Cui, K., & Zhao, K. (2013). Intragenic DNA methylation

modulates alternative splicing by recruiting MeCP2 to promote exon recognition. Cell

Res, 23(11), 1256-1269. doi: 10.1038/cr.2013.110

McCabe, M. J., Gaston-Massuet, C., Tziaferi, V., Gregory, L. C., Alatzoglou, K. S., Signore, M.,

. . . Dattani, M. T. (2011). Novel FGF8 mutations associated with recessive

holoprosencephaly, craniofacial defects, and hypothalamo-pituitary dysfunction. J Clin

Endocrinol Metab, 96(10), E1709-1718. doi: 10.1210/jc.2011-0454

Mellon, P. L., Windle, J. J., Goldsmith, P. C., Padula, C. A., Roberts, J. L., & Weiner, R. I.

(1990). Immortalization of hypothalamic GnRH neurons by genetically targeted

tumorigenesis. Neuron, 5(1), 1-10.

Mendonca, A., Chang, E. H., Liu, W., & Yuan, C. (2014). Hydroxymethylation of DNA

influences nucleosomal conformation and stability in vitro. Biochim Biophys Acta,

1839(11), 1323-1329. doi: 10.1016/j.bbagrm.2014.09.014

Neri, F., Incarnato, D., Krepelova, A., Rapelli, S., Pagnani, A., Zecchina, R., . . . Oliviero, S.

(2013). Genome-wide analysis identifies a functional association of Tet1 and Polycomb

repressive complex 2 in mouse embryonic stem cells. Genome Biol, 14(8), R91. doi:

10.1186/gb-2013-14-8-r91

Ngo, T. T., Yoo, J., Dai, Q., Zhang, Q., He, C., Aksimentiev, A., & Ha, T. (2016). Effects of

cytosine modifications on DNA flexibility and nucleosome mechanical stability. Nat

Commun, 7, 10813. doi: 10.1038/ncomms10813

O'Carroll, D., Erhardt, S., Pagani, M., Barton, S. C., Surani, M. A., & Jenuwein, T. (2001). The

polycomb-group gene Ezh2 is required for early mouse development. Mol Cell Biol,

21(13), 4330-4336. doi: 10.1128/mcb.21.13.4330-4336.2001

134

Ornitz, D. M. (2000). FGFs, heparan sulfate and FGFRs: complex interactions essential for

development. Bioessays, 22(2), 108-112. doi: 10.1002/(sici)1521-

1878(200002)22:2<108::aid-bies2>3.0.co;2-m

Santiago, M., Antunes, C., Guedes, M., Sousa, N., & Marques, C. J. (2014). TET enzymes and

DNA hydroxymethylation in neural development and function - how critical are they?

Genomics, 104(5), 334-340. doi: 10.1016/j.ygeno.2014.08.018

Schor, I. E., Allo, M., & Kornblihtt, A. R. (2010). Intragenic chromatin modifications: a new

layer in alternative splicing regulation. Epigenetics, 5(3), 174-179.

Schwanzel-Fukuda, M., & Pfaff, D. W. (1989). Origin of luteinizing hormone-releasing hormone

neurons. Nature, 338(6211), 161-164. doi: 10.1038/338161a0

Stevenson, E. L., Corella, K. M., & Chung, W. C. J. (2013). Ontogenesis of Gonadotropin-

Releasing Hormone Neurons: A Model for Hypothalamic Neuroendocrine Cell

Development. Front Endocrinol (Lausanne), 4, 89. doi: 10.3389/fendo.2013.00089

Stewart, C. E., Corella, K. M., Samberg, B. D., Jones, P. T., Linscott, M. L., & Chung, W. C. J.

(2016). Perinatal midline astrocyte development is impaired in fibroblast growth factor 8

hypomorphic mice. Brain Res, 1646, 287-296. doi: 10.1016/j.brainres.2016.06.015

Storm, E. E., Garel, S., Borello, U., Hebert, J. M., Martinez, S., McConnell, S. K., . . .

Rubenstein, J. L. (2006). Dose-dependent functions of Fgf8 in regulating telencephalic

patterning centers. Development, 133(9), 1831-1844. doi: 10.1242/dev.02324

Sui, X., Price, C., Li, Z., & Chen, J. (2012). Crosstalk Between DNA and Histones: Tet's New

Role in Embryonic Stem Cells. Curr Genomics, 13(8), 603-608. doi:

10.2174/138920212803759730

135

Sun, W., Zang, L., Shu, Q., & Li, X. (2014). From development to diseases: the role of 5hmC in

brain. Genomics, 104(5), 347-351. doi: 10.1016/j.ygeno.2014.08.021

Sun, X., Meyers, E. N., Lewandoski, M., & Martin, G. R. (1999). Targeted disruption of Fgf8

causes failure of cell migration in the gastrulating mouse embryo. Genes Dev, 13(14),

1834-1846.

Sunmonu, N. A., Li, K., Guo, Q., & Li, J. Y. (2011). Gbx2 and Fgf8 are sequentially required for

formation of the midbrain-hindbrain compartment boundary. Development, 138(4), 725-

734. doi: 10.1242/dev.055665

Suzuki-Hirano, A., & Shimogori, T. (2009). The role of Fgf8 in telencephalic and diencephalic

patterning. Semin Cell Dev Biol, 20(6), 719-725. doi: 10.1016/j.semcdb.2009.04.002

Szulwach, K. E., Li, X., Li, Y., Song, C. X., Wu, H., Dai, Q., . . . Jin, P. (2011). 5-hmC-mediated

epigenetic dynamics during postnatal neurodevelopment and aging. Nat Neurosci, 14(12),

1607-1616. doi: 10.1038/nn.2959

Teif, V. B., Beshnova, D. A., Vainshtein, Y., Marth, C., Mallm, J. P., Hofer, T., & Rippe, K.

(2014). Nucleosome repositioning links DNA (de)methylation and differential CTCF

binding during stem cell development. Genome Res, 24(8), 1285-1295. doi:

10.1101/gr.164418.113

Theil, T., Dominguez-Frutos, E., & Schimmang, T. (2008). Differential requirements for Fgf3

and Fgf8 during mouse forebrain development. Dev Dyn, 237(11), 3417-3423. doi:

10.1002/dvdy.21765

Toyoda, R., Assimacopoulos, S., Wilcoxon, J., Taylor, A., Feldman, P., Suzuki-Hirano, A., . . .

Grove, E. A. (2010). FGF8 acts as a classic diffusible morphogen to pattern the

neocortex. Development, 137(20), 3439-3448. doi: 10.1242/dev.055392

136

Tsai, P.-S., Brooks, L. R., Rochester, J. R., Kavanaugh, S. I., & Chung, W. C. J. (2011).

Fibroblast growth factor signaling in the developing neuroendocrine hypothalamus. Front

Neuroendocrinol, 32(1), 95-107. doi: https://doi.org/10.1016/j.yfrne.2010.11.002

Verma, N., Pan, H., Dore, L. C., Shukla, A., Li, Q. V., Pelham-Webb, B., . . . Huangfu, D.

(2018). TET proteins safeguard bivalent promoters from de novo methylation in human

embryonic stem cells. Nat Genet, 50(1), 83-95. doi: 10.1038/s41588-017-0002-y

Walshe, J., & Mason, I. (2003). Unique and combinatorial functions of Fgf3 and Fgf8 during

zebrafish forebrain development. Development, 130(18), 4337-4349.

Wray, S., Grant, P., & Gainer, H. (1989). Evidence that cells expressing luteinizing hormone-

releasing hormone mRNA in the mouse are derived from progenitor cells in the olfactory

placode. Proc Natl Acad Sci U S A, 86(20), 8132-8136.

Wu, H., D'Alessio, A. C., Ito, S., Xia, K., Wang, Z., Cui, K., . . . Zhang, Y. (2011). Dual

functions of Tet1 in transcriptional regulation in mouse embryonic stem cells. Nature,

473(7347), 389-393. doi: 10.1038/nature09934

Wu, H., & Zhang, Y. (2011). Mechanisms and functions of Tet protein-mediated 5-

methylcytosine oxidation. Genes Dev, 25(23), 2436-2452. doi: 10.1101/gad.179184.111

Wu, X., & Zhang, Y. (2017). TET-mediated active DNA demethylation: mechanism, function

and beyond. Nat Rev Genet, 18(9), 517-534. doi: 10.1038/nrg.2017.33

Xin, Y. J., Yuan, B., Yu, B., Wang, Y. Q., Wu, J. J., Zhou, W. H., & Qiu, Z. (2015). Tet1-

mediated DNA demethylation regulates neuronal cell death induced by oxidative stress.

Sci Rep, 5, 7645. doi: 10.1038/srep07645

137

Xu, J., Liu, Z., & Ornitz, D. M. (2000). Temporal and spatial gradients of Fgf8 and Fgf17

regulate proliferation and differentiation of midline cerebellar structures. Development,

127(9), 1833-1843.

Yosefzon, Y., David, C., Tsukerman, A., Pnueli, L., Qiao, S., Boehm, U., & Melamed, P. (2017).

An epigenetic switch repressing Tet1 in gonadotropes activates the reproductive axis.

Proc Natl Acad Sci U S A, 114(38), 10131-10136. doi: 10.1073/pnas.1704393114

Zemke, M., Draganova, K., Klug, A., Scholer, A., Zurkirchen, L., Gay, M. H., . . . Sommer, L.

(2015). Loss of Ezh2 promotes a midbrain-to-forebrain identity switch by direct gene

derepression and Wnt-dependent regulation. BMC Biol, 13, 103. doi: 10.1186/s12915-

015-0210-9

Zhang, Y., & Reinberg, D. (2001). Transcription regulation by histone methylation: interplay

between different covalent modifications of the core histone tails. Genes Dev, 15(18),

2343-2360. doi: 10.1101/gad.927301

Zhao, J., Ohsumi, T. K., Kung, J. T., Ogawa, Y., Grau, D. J., Sarma, K., . . . Lee, J. T. (2010).

Genome-wide identification of polycomb-associated RNAs by RIP-seq. Mol Cell, 40(6),

939-953. doi: 10.1016/j.molcel.2010.12.011

Zhou, Y., & Hu, Z. (2015). Genome-wide demethylation by 5-aza-2'-deoxycytidine alters the

cell fate of stem/progenitor cells. Stem Cell Rev, 11(1), 87-95. doi: 10.1007/s12015-014-

9542-z

138

CHAPTER V TET1 NULL MUTATION IN MICE DISRUPTS EMBRYONIC DEVELOPMENT OF GNRH NEURONS AND THEIR PHYSIOLOGICAL FUNCTION IN ADULTHOOD

INTRODUCTION

Epigenetic DNA modifications at CpG dinucleotides in mammalian cells are known to be critical for controlling gene transcription. (Antunes et al., 2019; Ross & Bogdanovic,

2019; Santiago et al., 2014; Sun et al., 2014; Wu & Zhang, 2017). In this context, we demonstrated that epigenetic factors that modify DNA methylation levels on the promoter of fibroblast growth factor (FGF) 8 were important for the transcriptional activity of Fgf8 during the embryonic development of the mouse olfactory placode (OP)

(Megan L. Linscott & Wilson CJ Chung, 2016; Linscott & Chung, 2019; Schwanzel-

Fukuda et al., 1989; Wray, Grant, et al., 1989). These observations are of physiological relevance because we previously showed that FGF8 function in the embryonic OP is required for the ontogenesis of gonadotropin-releasing hormone (GnRH) neurons, which are known to confer fertility in vertebrates, including human and non-human primates, rodents and birds (Chung et al., 2010; W. C. Chung et al., 2008; Falardeau et al., 2008; P.

S. Tsai et al., 2011).

139

Further analysis of the mouse embryonic OP Fgf8 promoter revealed that it undergoes rapid demethylation between embryonic days (E) 9.5 and E13.5, which coincides with GnRH neuron ontogenesis. DNA demethylation, which is associated with gene activation, is catalyzed by the family of Ten-eleven translocase (TET) enzymes, consisting of three members, which convert methylated cytosines (5mC) to hydroxymethylated cytosines (5hmC) (Branco et al.,

2011; Wu & Zhang, 2017). TET proteins can further convert 5hmC to 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC), which can be replaced by base excision repair mechanisms with an unmethylated cytosine (Kohli & Zhang, 2013). Indeed, our analysis of the Fgf8 promoter found that it was highly enriched for 5hmC rather than 5mC in the early embryonic OP. In addition, further studies showed that genome-wide mouse OP Tet mRNA expression increased significantly over this period of time, and that TET1 binding to the Fgf8 promoter is important for Fgf8 transcription activity (Linscott & Chung, 2019). Taken together, these data led us to conclude that Fgf8 transcriptional rate is a TET-dependent process.

Interestingly, in females, Tet1 knockout mice have abnormal ovarian follicle development with markedly smaller ovaries and fewer mature follicles (Dawlaty et al., 2013;

Dawlaty et al., 2011). Although, these data suggest that there is a complex interplay between hormonal regulation and DNA demethylation programs regulated by the TET proteins, the underlying cause(s) are not well understood. One possibility is that inactivation of TET1 function in mice may disrupt OP Fgf8 transcription, and consequently result in abnormal GnRH neuron ontogenesis. To test this question, we investigated the role of TET1 in the developing mouse hypothalamus. Specifically, we measured neurogenic gene expression in wildtype (WT) versus

Tet1 (Tet1-/-) knockout animals. We also measured the reproductive capacity of mice harboring this mutation using several fertility assays. Fertility assays were followed by

140 immunohistochemistry in PN0 females for GnRH immunoreactivity. Finally, we profiled mRNA changes in GnRH in the E13.5 OP, during GnRH neuron emergence. Overall, we found that

TET1 is compatible with reproduction; however, our studies suggest Tet1-/- mice have a reduced ability to participate in several neuronal development processes.

MATERIALS AND METHODS

Tet1 knockout mice timed breeding and nasal explant tissue

Adult heterozygous C57BL/6 x 129S4/SvJae male Tet1+/- x female Tet1+/- mice were timed-bred in the late afternoon in our animal facility (12L:12D cycle) with access to food and water ad libitum. All procedures were approved by the Institutional Animal Care and Use Committee at

Kent State University. In the morning, females with a sperm plug were denoted as embryonic day (E) 0.5. Adult pregnant C57BL/6 x 129S4/SvJae Tet1+/- female mice were sacrificed at

E10.5 or E13.5. The uterine horns were quickly removed from the mice and kept in sterile ice- cold phosphate-buffered saline (Sigma-Aldrich, P3813). Following, the nasal region containing the OPs was surgically isolated using a dissection microscope. For expression assays of E10.5, or

E13.5 nasal tissues were collected, immediately flash frozen and kept at -80C.

RT-qPCR

Strand cDNA Synthesis Kit (New England Biolabs, E6560L). ProtoScript® II First Strand cDNA Synthesis Kit was used to reverse transcribe 0.5 µg of total RNA. RT-qPCR was

141 performed in triplicate with gene-specific, intron-spanning primers using a Mastercycler EP

Realplex2 (Eppendorf, EPPE6300000.604) with SYBR Green PCR Master Mix (Roche,

04707516001). Relative mRNA expression levels were calculated using the ΔΔ-2CT method

(Livak & Schmittgen, 2001). Hypoxanthine phosphoribosyltransferase 1 (Hprt-1) was used as a housekeeping gene.

Brain tissue collection

PN0, WT, and Tet1-/- brain tissue was obtained by euthanizing newborn pups within hours after birth. Adult WT and Tet1+/- brain tissue were obtained from 2 to 4-month-old female mice. The brains were immersion-fixed in 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer, stored in 30% sucrose, and genotyped using PCR for TET1 and SRY. Serial coronal WT, Tet1+/-, and Tet1-/- sections (25 µm) were obtained using a cryostat (Leica CM 1950, Buffalo Grove, IL), and thaw-mounted on slides coated with gelatin (Sigma-Aldrich, St. Louis, MO).

Immunohistochemistry

Sections from WT, Tet1+/- and Tet1-/- mice were processed simultaneously to minimize the variability in immunoreactivity. Serial coronal sections were incubated in 1% hydrogen peroxide/TBS solution for 15 min at room temperature, washed in TBS, 3×5 min, and incubated in primary rabbit polyclonal anti-GnRH (1:15,000), made in TBS/0.3% Triton-X (Fisher

Scientific, Pittsburgh, PA) and 2% normal goat serum for 2 days at 4 °C. Sections were washed and incubated with biotinylated-goat anti-rabbit (1:600) for 2 h at room temperature followed by

ABC (1:800) (Vector Laboratories, Burlingame, CA) in TBS for 2 h at room temperature, and reacted with 0.05% diaminobenzidine/0.1% nickel ammonium sulfate (Sigma-Aldrich, St. Louis,

142

MO) and 0.01% H2O2/TBS for 20 min. Sections were dehydrated with ethanol, cleared with histoclear and cover slipped with permount.

Image analysis

For our quantification studies, we analyzed the immunoreactivity of GnRH peptide in matched median eminence images. Grayscale digital images were captured using a 10× objective mounted on an Olympus microscope fitted with a color camera (SC30, Olympus, Corporation of the

Americas, Center Valley, PA) connected to a PC computer. GnRH immunoreactivity was analyzed using Olympus CellSens software (Olympus Corporation of the Americas, Center

Valley, PA). A threshold mask was generated which reliably and accurately covered the GnRH immunoreactivity. Total GnRH peptide density was measured as immunoreactivity covered by pixels in a fixed region of interest rectangle 4,858 µm2).

PCR Array

Total RNA was isolated, and reverse transcribed as listed above. The cDNA was used on the real-time RT² Profiler PCR Array (QIAGEN, Cat. no. PAHS-404Z) in combination with Roche,

04707516001. CT values were exported to an Excel file to create a table of CT values. This table was then uploaded on to the data analysis web portal at http://www.qiagen.com/geneglobe.

Samples were assigned to controls and test groups. CT values were normalized based on a/an

Automatic Selection from Full Panel of reference genes. The data analysis web portal calculates fold change/regulation using delta delta CT method, in which delta CT is calculated between gene of interest (GOI) and an average of reference genes (HKG), followed by delta-delta CT calculations (delta CT (Test Group)-delta CT (Control Group)). Fold Change was then calculated

143 using 2^ (-delta delta CT) formula. The data analysis web portal also plots scatter plot, volcano plot, clustergram, and heat map. This data analysis report was exported from the QIAGEN web portal at GeneGlobe.

Statistical analysis

0.05.

RESULTS

Neurogenesis array in Tet1-/- E13.5 OPs

Many previous studies have elucidated the role of TET1 in neurogenesis and in developmental processes. Additionally, our previous studies found that 5hmC levels rise during OP development as well as Tet mRNA, indicating a major role in the developing OP (Linscott &

Chung, 2019). Here, we found that in the E13.5 OP, Tet1-/- animals have markedly different expression patterns of key neurogenesis regulators. Specifically, we found that PAX3 and PAX6 were significantly downregulated in Tet1-/- OPs (p < 0.05). We also found that CDK5R1was upregulated in Tet1-/- mice (p < 0.05).

144

Figure 5.1. Heat map of the E13.5 mouse OP in WT versus Tet1-/-. 2-fold gene expression patterns of upregulated genes in red, and downregulated genes represented in green.

145

Reproductive capacity of Tet1-/- mice

Previous studies have indicated that Tet1 knockout mice have abnormal follicular development, therefore we tested the reproductive capacity of Tet1-/- mice. First, we compared the weights of male and female pups at PN0. We found a significant difference in the weight of female Tet1-/- animals, compared to female WT littermates (Fig. 5.1A; p <0.05). Interestingly, while no differences were found in exclusively male or female animals, Tet1+/- x Tet1+/- mating resulted in a significant reduction in the number of pups per litter produced (Fig. 5.2D; p < 0.05). However, there was no significant differences in the fertility index, where number of litters within a 90-day period were measured (Fig. 5.2C), or in days until first litter was produced (Fig. 5.2B).

146

Figure 5.2. Reproductive capacity of Tet1-/- mice. A) Weight in grams (g) of male or female

WT, Tet1+/-, or Tet1-/- animals at birth. Females displayed a significantly lower birthweight than their WT littermates (p < 0.05; n = 10-20). B) Number of days to produce litter after pairing with male. C) Number of litters produced over a 90-day period. D) Number of pups per litter produced. Male Tet1+/- x female Tet1+/- pairing produces significantly less pups than WT x WT pairing (p < 0.05; n = 4).

147

GnRH immunoreactivity in female PN0 TET1 KO mice

Reproductive competency is dependent on production of GnRH. Our previous studies found that

Tet1 reduction could significantly downregulate Fgf8 mRNA production in GT1-7 neurons.

Here, we hypothesized that in early GnRH neuron development, TET1 reduction could alter Fgf8 transcriptional timing and therefore disrupt GnRH neurogenesis. We also found that female

Tet1+/- mice has lower birthweights than their WT littermates. We therefore tested the immunoreactivity of GnRH neurons in the median eminence of WT or Tet1-/- female PN0 animals. Interestingly, we found a significant decrease in GnRH immunoreactivity in Tet1-/- animals (p < 0.05; n = 4).

148

Figure 5.3. Immunoreactivity of GnRH in female WT or Tet1-/- PN0 mice. Representative photographs of the median eminence of A) WT or B) Tet1-/-. C) Quantification of immunoreactivity in µm2. Significantly less GnRH immunoreactivity in Tet1-/- mice (p < 0.05; n

= 4).

149

Expression of Fgf8 and GnRH mRNA in TET1-/- mouse OPs

Our previously published data indicated that TET1 binds the Fgf8 promoter in the OP and transcriptionally regulates expression of Fgf8 and Fgfr1 mRNA in GT1-7 neurons. To further assess the ability of TET1 to regulate gene transcription of Fgf8 in vivo, we used Tet1 knockout mice to measure Fgf8, GnRH and Fgfr1 mRNA production in the OP. We found that Tet1-/- mice exhibit a decrease in Fgf8 and GnRH mRNA (p < 0.05; n = 3) in the E13.5 mouse OP compared to WT littermates (Figure 5.4). However, Fgfr1 expression was not statistically significant (p =

0.07).

150

Figure 5.4. Fgf8, Fgfr1, and GnRH mRNA expression in the E13.5 OP of WT versus Tet1-/- animals. Asterisk is p < 0.05 (n = 3).

151

DISCUSSION

Timing of gene expression can be explained, in part, by epigenetic events which control gene expression in a timely and coordinated manner. Here, we showed that TET1 regulates key neurogenic genes in the developing OP. We also found that loss of TET1 can reduce several measures of fertility. These results were substantiated when female PN0 Tet1-/- mice demonstrated a significant reduction in GnRH immunoreactivity. Finally, we showed that during

GnRH neuron emergence, GnRH mRNA is reduced in the E13.5 OP of Tet1-/- animals. These results underscore the importance of demethylation, specifically via TET1, during embryonic

GnRH neuron development.

Previous studies have demonstrated the role of TET1 in gene expression control of neurogenesis and in the neuroendocrine system. For example, TET1 overexpression upregulates several neuronal markers during fetal brain development, is responsive to hormonal feedback from gonadal steroids, and has recently been shown to be involved in sexual differentiation postnatally (Cisternas et al., 2020; Kim et al., 2016; Yosefzon et al., 2017). However, little is known about the role of TET1 in the developing neuroendocrine system. To better address this, we profiled mRNA expression of genes involved in neurogenesis pathways in the developing OP during GnRH neuron onset. Interestingly, we found that loss of TET1 caused a reduction in several critical neurogenesis genes, such as Pax3 and Pax6, and an increase in Cdk5r1.

The family of PAX transcription factors are well studied in neural development, especially in neural crest cell lineage and OP cell fate (Bhattacharyya & Bronner-Fraser, 2008;

Monsoro-Burq, 2015). Interestingly, point mutations in Pax6 have been associated with isolated hypogonadotropic hypogonadism, specifically in the Chinese male population (Zhou et al.,

2018). Further studies have found that “small eye” Pax6 null mice fail to develop eyes and an

152 olfactory placode, with virtually no GnRH immunoreactivity in any region of the brain (Hill et al., 1992; Skynner et al., 1999). We also found that Cdk5r1, a gene involved in neuronal migration and intellectual disability, was upregulated 2-fold in Tet1-/- mice (Moncini et al., 2016;

Moncini et al., 2011).

We previously demonstrated Tet1, 2 and 3 mRNA levels significantly increased in the mouse OP. Also, TET1 was bound to the embryonic mouse OP the Fgf8 promoter. These cellular changes coincided with 5hmC enrichment of the Fgf8 promoter and enhance Fgf8 transcription.

(Linscott & Chung, 2019). Therefore, we predicted that embryonic OP Fgf8 transcription is abnormal in Tet1 null mice. We also predicted disrupted embryonic GnRH neuron development and function due to FGF8’s importance with regard to the development of the GnRH neuronal system in the embryonic mouse OP (Chung et al., 2010; W. C. Chung et al., 2008; Falardeau et al., 2008). Here, we demonstrate that early development of GnRH neurons requires embryonic

TET1 function. Specifically, we showed that both Fgf8 and GnRH mRNA are significantly downregulated in the E13.5 mouse OP when TET1 function is eliminated. Moreover, these results were supported by our findings showing that female PN0 Tet1-/- mice demonstrated a significant reduction in GnRH immunoreactivity when compared to WT littermates. These data led to the conclusion that elimination of TET1 function alone was sufficient in disrupting FGF8 and GnRH neuron development and function. Currently, we do not know whether the reduction in GnRH expression is due to direct TET1 effects on GnRH expression or due to its effects on the embryonic mouse OP Fgf8 expression.

In conclusion, these studies indicate that environmental and hormonal influences on the epigenome provide an additional layer of complexity to developmental programs which require precise timing of gene expression. In the present study, we demonstrate that TET1 is required for

153

GnRH neuronal development, and effects transcription of key genes required for GnRH neuronal onset. Overall, these studies suggest that demethylation via TET1 controls development of neuroendocrine genes, which are important for early GnRH neuronal fate-specification and migration.

154

REFERENCES

Antunes, C., Sousa, N., Pinto, L., & Marques, C. J. (2019). TET enzymes in neurophysiology

and brain function. Neurosci Biobehav Rev, 102, 337-344. doi:

10.1016/j.neubiorev.2019.05.006

Bhattacharyya, S., & Bronner-Fraser, M. (2008). Competence, specification and commitment to

an olfactory placode fate. Development, 135(24), 4165-4177. doi: 10.1242/dev.026633

Branco, M. R., Ficz, G., & Reik, W. (2011). Uncovering the role of 5-hydroxymethylcytosine in

the epigenome. Nat Rev Genet, 13(1), 7-13. doi: 10.1038/nrg3080

Chung, W. C., Matthews, T. A., Tata, B. K., & Tsai, P. S. (2010). Compound deficiencies in

multiple fibroblast growth factor signalling components differentially impact the murine

gonadotrophin-releasing hormone system. J Neuroendocrinol, 22(8), 944-950. doi:

10.1111/j.1365-2826.2010.02024.x

Chung, W. C., Moyle, S. S., & Tsai, P. S. (2008). Fibroblast growth factor 8 signaling through

fibroblast growth factor receptor 1 is required for the emergence of gonadotropin-

releasing hormone neurons. Endocrinology, 149(10), 4997-5003. doi: 10.1210/en.2007-

1634

Cisternas, C. D., Cortes, L. R., Bruggeman, E. C., Yao, B., & Forger, N. G. (2020).

Developmental changes and sex differences in DNA methylation and demethylation in

hypothalamic regions of the mouse brain. Epigenetics, 15(1-2), 72-84. doi:

10.1080/15592294.2019.1649528

Dawlaty, M. M., Breiling, A., Le, T., Raddatz, G., Barrasa, M. I., Cheng, A. W., . . . Jaenisch, R.

(2013). Combined deficiency of Tet1 and Tet2 causes epigenetic abnormalities but is

155

compatible with postnatal development. Dev Cell, 24(3), 310-323. doi:

10.1016/j.devcel.2012.12.015

Dawlaty, M. M., Ganz, K., Powell, B. E., Hu, Y. C., Markoulaki, S., Cheng, A. W., . . . Jaenisch,

R. (2011). Tet1 is dispensable for maintaining pluripotency and its loss is compatible

with embryonic and postnatal development. Cell Stem Cell, 9(2), 166-175. doi:

10.1016/j.stem.2011.07.010

Falardeau, J., Chung, W. C., Beenken, A., Raivio, T., Plummer, L., Sidis, Y., . . . Pitteloud, N.

(2008). Decreased FGF8 signaling causes deficiency of gonadotropin-releasing hormone

in humans and mice. J Clin Invest, 118(8), 2822-2831. doi: 10.1172/jci34538

Hill, R. E., Favor, J., Hogan, B. L., Ton, C. C., Saunders, G. F., Hanson, I. M., . . . van

Heyningen, V. (1992). Mouse Small eye results from mutations in a paired-like

homeobox-containing gene. Nature, 355(6362), 750. doi: 10.1038/355750a0

Kim, H., Jang, W. Y., Kang, M. C., Jeong, J., Choi, M., Sung, Y., . . . Ryoo, Z. Y. (2016). TET1

contributes to neurogenesis onset time during fetal brain development in mice. Biochem

Biophys Res Commun, 471(4), 437-443. doi: 10.1016/j.bbrc.2016.02.060

Kohli, R. M., & Zhang, Y. (2013). TET enzymes, TDG and the dynamics of DNA

demethylation. Nature, 502(7472), 472-479. doi: 10.1038/nature12750

Linscott, M. L., & Chung, W. C. (2016). Fibroblast Growth Factor 8 Expression in GT1-7

GnRH-Secreting Neurons is Androgen-Independent, but can be Upregulated by the

Inhibition of DNA Methyltransferases. Frontiers in Cell and Developmental Biology, 4.

doi: 10.3389/fcell.2016.00034

156

Linscott, M. L., & Chung, W. C. J. (2019). TET1 regulates fibroblast growth factor 8

transcription in gonadotropin releasing hormone neurons. PLoS One, 14(7), e0220530.

doi: 10.1371/journal.pone.0220530

Livak, K. J., & Schmittgen, T. D. (2001). Analysis of relative gene expression data using real-

time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods, 25(4), 402-408.

doi: 10.1006/meth.2001.1262

Moncini, S., Castronovo, P., Murgia, A., Russo, S., Bedeschi, M. F., Lunghi, M., . . . Venturin,

M. (2016). Functional characterization of CDK5 and CDK5R1 mutations identified in

patients with non-syndromic intellectual disability. J Hum Genet, 61(4), 283-293. doi:

10.1038/jhg.2015.144

Moncini, S., Salvi, A., Zuccotti, P., Viero, G., Quattrone, A., Barlati, S., . . . Riva, P. (2011). The

role of miR-103 and miR-107 in regulation of CDK5R1 expression and in cellular

migration. PLoS One, 6(5), e20038. doi: 10.1371/journal.pone.0020038

Monsoro-Burq, A. H. (2015). PAX transcription factors in neural crest development. Semin Cell

Dev Biol, 44, 87-96. doi: 10.1016/j.semcdb.2015.09.015

Ross, S. E., & Bogdanovic, O. (2019). TET enzymes, DNA demethylation and pluripotency.

Biochem Soc Trans, 47(3), 875-885. doi: 10.1042/bst20180606

Santiago, M., Antunes, C., Guedes, M., Sousa, N., & Marques, C. J. (2014). TET enzymes and

DNA hydroxymethylation in neural development and function - how critical are they?

Genomics, 104(5), 334-340. doi: 10.1016/j.ygeno.2014.08.018

Schwanzel-Fukuda, M., Bick, D., & Pfaff, D. W. (1989). Luteinizing hormone-releasing

hormone (LHRH)-expressing cells do not migrate normally in an inherited hypogonadal

(Kallmann) syndrome. Mol Brain Res, 6(4), 311-326.

157

Skynner, M. J., Slater, R., Sim, J. A., Allen, N. D., & Herbison, A. E. (1999). Promoter

transgenics reveal multiple gonadotropin-releasing hormone-I-expressing cell populations

of different embryological origin in mouse brain. J Neurosci, 19(14), 5955-5966.

Sun, W., Zang, L., Shu, Q., & Li, X. (2014). From development to diseases: the role of 5hmC in

brain. Genomics, 104(5), 347-351. doi: 10.1016/j.ygeno.2014.08.021

Tsai, P. S., Brooks, L. R., Rochester, J. R., Kavanaugh, S. I., & Chung, W. C. (2011). Fibroblast

growth factor signaling in the developing neuroendocrine hypothalamus. Front

Neuroendocrinol, 32(1), 95-107. doi: 10.1016/j.yfrne.2010.11.002

Wray, S., Grant, P., & Gainer, H. (1989). Evidence that cells expressing luteinizing hormone-

releasing hormone mRNA in the mouse are derived from progenitor cells in the olfactory

placode. Proc Natl Acad Sci U S A, 86(20), 8132-8136.

Wu, X., & Zhang, Y. (2017). TET-mediated active DNA demethylation: mechanism, function

and beyond. Nat Rev Genet, 18(9), 517-534. doi: 10.1038/nrg.2017.33

Yosefzon, Y., David, C., Tsukerman, A., Pnueli, L., Qiao, S., Boehm, U., & Melamed, P. (2017).

An epigenetic switch repressing Tet1 in gonadotropes activates the reproductive axis.

Proc Natl Acad Sci U S A, 114(38), 10131-10136. doi: 10.1073/pnas.1704393114

Zhou, C., Niu, Y., Xu, H., Li, Z., Wang, T., Yang, W., . . . Liu, J. (2018). Mutation profiles and

clinical characteristics of Chinese males with isolated hypogonadotropic hypogonadism.

Fertil Steril, 110(3), 486-495 e485. doi: 10.1016/j.fertnstert.2018.04.010

158

CHAPTER VI GENERAL DISCUSSION

Neuronal development is dependent on proper timing of gene expression which coordinate morphogen gradients, transcription factor networks and signaling cascades.

FGF signaling during early neuronal development is responsible for generating multiple brain regions. FGF8 and FGFR1 deficiency severely compromises vertebrate reproduction in mice and humans and causes KS (W. C. Chung et al., 2008; Falardeau et al., 2008). Our laboratory demonstrated that FGF8 signaling through FGFR1 is necessary for proper GnRH neuron development in mice (Chung et al., 2010; Chung & Tsai, 2010;

W. C. J. Chung et al., 2008; Falardeau et al., 2008). Perturbations in this signaling system disrupt the ontogenesis and physiological function of GnRH neurons.

Here, we examined the role of Fgf8 transcription during GnRH neuron onset.

Previous studies have focused on FGF8 in the context of ligand-receptor signaling, which is required for GnRH neuron onset in humans. Similarly, in mouse models, homozygous mutations in Fgf8 result in the absence of GnRH neurons, as well as missing olfactory bulbs and hypothalamic structures (Sato et al., 2017). Because Fgf8 is secreted during a narrow time window during GnRH neuron onset, we sought to understand the transcriptional mechanisms which drive Fgf8 mRNA production and how they contribute to GnRH neuron onset. We also sought to understand the epigenomic contribution on both Fgf8, and on the OP genome. Specifically, to study how the epigenome contributes

159 to GnRH neuronal development, we developed an embryonic explant model, utilizing the mouse

OP, the birthplace of GnRH neurons.

We tested several previously identified transcription factors and their ability to interact with the promoter region of Fgf8 or influence transcriptional rates. We found that ARs were constitutively bound to the promoter region of Fgf8. Surprisingly, activation of ARs with its cognate ligand DHT, did not affect Fgf8 transcription in the embryonic mouse OP (M. L.

Linscott & W. C. Chung, 2016). Interestingly, by inducing a demethylated state, Fgf8 mRNA increased, which could be prevented by DHT treatment (M. L. Linscott & W. C. Chung, 2016).

These results suggested that DHT activated ARs may act as a quencher of Fgf8 transcription rather than the hypothesized activational function.

Nevertheless, these data led to the observation that Fgf8 transcription is under the influence of epigenetic factors that control DNA methylation levels. Therefore, we performed

MeDIP experiments in the mouse OP, and found that Fgf8 promoter is demethylated during

GnRH neuron onset (Linscott & Chung, 2019). Accordingly, we directed our further studies towards determining the role of demethylation on Fgf8 transcription during GnRH ontogenesis.

Our immunoprecipitation studies demonstrate that Fgf8 in the E9.5 OP is enriched with both

5hmC and TET1 at specific CpG islands, which decreased with age (Linscott & Chung, 2019).

TET1 has also been shown to play a direct role in maintaining histone bivalency (Kong et al.,

2016; Voigt et al., 2013). Indeed, we found that Fgf8 expression is controlled by both bivalent histone modifications and demethylation events. Two modifications, H3K27me3 (a repressive histone) and H3K4me3 (an activating histone), form bivalent domains that regulate the chromatin compaction status of genes, and are thought to respond to developmental cues.

Specifically, we found that in the E9.5 OP, Fgf8 harbors both H3K4me3 and H3K27me3, while

160 at E13.5, only H3K27me3 is present, indicating a possible role for Fgf8-dependent histone modifications in specifying GnRH neuron fate (Linscott & Chung, 2019). Recently, we have shown through siRNA experiments that both Fgf8 and Fgfr1, are regulated by TET1, indicating

TET1 is an important upstream epigenetic factor involved in the regulation of early GnRH neurons (Linscott & Chung, 2019).

AR/RAR control of Fgf8 transcription

In our initial studies, we hypothesized that transcriptional activation of Fgf8 depended on recruitment of several important transcription factors. Fgf8 harbors many consensus sequences throughout the promoter region located near the TSS and at distal sites. Previous studies have determined that these sites are able to recruit ARs at 3 specific AREs as well as RARs at a distally located RARE site (Gnanapragasam et al., 2002; Kumar & Duester, 2014). Recruitment of these transcription factors was also demonstrated to significantly change expression patterns of Fgf8. We hypothesized that these previously identified transcription factors could also be recruited during early development and may regulate the transcriptional bursting event during early OP development. However, while RA is locally synthesized in the OP, RARs are not recruited to the Fgf8 promoter during OP development. On the other hand, we found that ARs were recruited to the Fgf8 promoter at all 3 AREs in adult GT1-7 neurons (M. L. Linscott & W.

C. Chung, 2016). Interestingly, AR binding was independent of DHT treatment, indicating that

ARs are constitutively bound to the Fgf8 promoter. However, DHT treatment did not affect Fgf8 transcription in either GT1-7 neurons or in developing OP cultures. We then hypothesized that

ARs may not exert an effect on Fgf8 transcription levels due to methylation status. To test this, we treated GT1-7 neurons with AZA and DHT and measured Fgf8 and GnRH expression.

161

Interestingly, we found that AZA treatment could significantly increase Fgf8 transcription levels, suggesting that Fgf8 is methylated (M. L. Linscott & W. C. Chung, 2016). When treated with both AZA and DHT however, DHT transcriptionally repressed Fgf8 expression (M. L. Linscott

& W. C. Chung, 2016). This suggested that while Fgf8 allows for AR binding on three separate

ARE consensus sites, the function of androgen signaling may not be to regulate Fgf8 transcription, but rather to moderate/inhibit the activational effects of other molecular processes, such as DNA methylation. These data called for a more detailed examination of epigenetic factors that could influence the developing GnRH neuron population.

Epigenetic control of OP Fgf8 transcription

Given that Fgf8 hypomorphy causes the absence of the GnRH neuronal system, we then focused on unraveling the precise control of Fgf8 transcriptional activity during embryonic mouse OP development. Transient Fgf8 mRNA expression coincides with the emergence phase of GnRH neuron development (Wray, Grant, et al., 1989). Following an activation period between ~E8.5 -

E10.5, Fgf8 transcription in the OP is rapidly repressed after E11.5 (Bachler & Neubuser, 2001).

Initially, we documented that Fgf8 mRNA expression in the developing OP is sensitive to changes in DNMT activity suggesting that epigenetic factors play important roles in the control of Fgf8 transcriptional activity (Megan L. Linscott & Wilson CJ Chung, 2016). When E10.5 mouse OPs or GT1-7 neurons were treated with AZA Fgf8 mRNA expression showed a significant increase (Megan L. Linscott & Wilson CJ Chung, 2016; Linscott & Chung, 2019).

Our subsequent studies found that Dnmt3a mRNA increased 10-fold, while Dnmt3b mRNA was reduced in the developing mouse OP during GnRH neuron emergence (Linscott & Chung, 2019).

To our surprise, neither DNMT1 nor DNMT3a were enriched on the OP Fgf8 promoter at E9.5

162 or E13.5, whereas we found that DNMT3b was enriched at E9.5 around the transcription start site (TSS), but not on E13.5. This led us to conclude that DNA methylation rate of the embryonic

OP Fgf8 promoter was likely reduced (Linscott & Chung, 2019). These studies did not directly measure Dnmt mRNA or protein expression in GnRH (precursor) neurons, and therefore their role during regarding GnRH development remain to be resolved.

These initial epigenetic findings formed the impetus for a detailed examination of the epigenetic landscape of the Fgf8 promoter in the embryonic mouse OP, which revealed that it harbors several CpG islands, and is sensitive to epigenetic factors (Linscott & Chung, 2019).

Interestingly, 5mC and 5hmC modifications were found to be significantly present in E8.5 OPs

(Linscott & Chung, 2019). In contrast, between E9.5 - E13.5, the OP Fgf8 promoter was only enriched for 5hmC, and not 5mC. Specifically, 5hmC was highly present in two CpG regions, one distal and one proximal to TSS of Fgf8. This apparent 5mC-to-5hmC enrichment of the Fgf8 promoter also coincided with an concurrent increase in Tet1-3 mRNA levels between E9.5 -

E13.5 (Linscott & Chung, 2019). These data indicate that TETs have a catalytically active function in converting 5mC-to-5hmC on the Fgf8 promoter and support our inference that increased Fgf8 transcription in the mouse OP is, in part, due to rapid TET-dependent demethylation of the Fgf8 promoter (Linscott & Chung, 2019). This inference was further supported by two lines of experimental evidence: 1) TET1 was bound to 5hmC-rich regions on the Fgf8 promoter (Linscott & Chung, 2019), and 2) Tet1 siRNA, which reduced Tet1 mRNA expression by ~75% in GT1-7 neurons, caused a 30% reduction in Fgf8 mRNA expression

(Linscott & Chung, 2019).

Our results indicate that TET1 directly controls Fgf8 expression and a reduction in TET1 could significantly reduce Fgf8 expression (Linscott & Chung, 2019). While we found no

163 difference in GnRH mRNA production, we posit that because the GT1-7 cell line models adult

GnRH neurons, that in a developmentally relevant model, Tet1 reduction may have a more powerful inhibitory effect on Fgf8 mRNA, which would prevent the onset of GnRH neurons.

Indeed, our Tet1-/- animals displayed a lower expression level of both Fgf8 and GnRH mRNA.

Moreover, female PN0 Tet1-/- mice had lower GnRH immunoreactivity in the median eminence.

Unpublished western blotting experiments in GN11 neurons showed that the addition of FGF8b, the most abundant Fgf8 isoform, increases phosphorylation of MAPK, but does not affect AKT phosphorylation. From these studies, we conclude that the role of Fgf8 in developing GnRH neurons is likely a differentiation or proliferation pathway rather than preventing apoptosis.

Therefore, TET1 regulation of Fgf8 during the E10.5 period is likely involved in a fate- specification of a GnRH neuronal population, which may not be present in adult GT1-7 neurons.

TET1 controls Fgf8 transcription via histone modifications

TET functions are not limited to its enzymatic ability to catalyze CpG demethylation (Wu et al.,

2011; Wu & Zhang, 2011). For example, TET1 can recruit many binding partners, including

EZH2, which further complicates its regulatory role as an epigenetic controller of gene transcription (Cartron et al., 2013; Wu et al., 2011; Wu & Zhang, 2011). EZH2 is known to elevate repressive H3K27me3 marks, which places a gene in a repressed state by promoting chromatin condensation (Kumar & Duester, 2014; Wang et al., 2017; Wu et al., 2011).

Here, we tested whether TET1 not only demethylates the Fgf8 promoter but could also recruit chromatin modifications. Indeed, concomitant ChIP pulldowns revealed EZH2 co-recruitment with TET1-bound sites on the Fgf8 promoter, and was enriched in E13.5 but not E9.5 mouse

OPs (Linscott & Chung, 2019). The presence of EZH2 on the TET1-associated Fgf8 promoter

164 regions overlaps with repressed Fgf8 transcription in a temporally dependent manner. Our studies confirmed that elevated H3K27me3 association with the mouse OP Fgf8 promoter coincided with EZH2 enrichment (Linscott & Chung, 2019). In contrast, the associated activational mark, H3K4me3, was enriched only on the E9.5 and not E13.5 Fgf8 promoter

(Linscott & Chung, 2019). Moreover, Ezh2 mRNA was significantly expressed at E10.5 compared to E13.5 in the mouse OP during the onset of GnRH neuron emergence (Linscott &

Chung, 2019). These data led to our working hypothesis postulates that the presence of TET1 on the Fgf8 promoter not only catalyzed its demethylation, but may also function as a molecular anchor for the recruitment of the PRC2 complex (Wu et al., 2011; Wu & Zhang, 2011). These data underscore the dynamics of the TET system during neuronal differentiation, and illustrate its dual effects (i.e., activation and repression) on the Fgf8 promoter to achieve transient OP Fgf8 transcription, and promote GnRH neuron development (Linscott & Chung, 2019; Wu et al.,

2011).

165

Figure 6.1. Early in development, TET1 demethylates the Fgf8 promoter and recruits histone bivalency markers, which activate Fgf8 transcription. After E11.5, EZH2 is recruited, which increases H3K27Me3 levels to condense chromatin and repress Fgf8 transcription.

166

The embryonic mouse OP undergoes significant DNA demethylation

Coordination of gene expression during neuronal fate-specification is tightly controlled process which can be accomplished at least partially through epigenetics. Interestingly, neuronal differentiation is correlated with an enrichment of 5hmC (Maria A. Hahn et al., 2013), which continues after birth, and may account for other developmental processes such as maturation, neuronal pathfinding, and synaptic transmission. While one study demonstrated that 5hmC levels are highest in hypothalamus (Münzel et al., 2010), we are the first, to our knowledge, to demonstrate that GnRH neurons also demonstrate genome-wide demethylation upon maturation.

Here, we investigated GnRH neuron emergence in the OP at E10.5 and E13.5 and found an upregulation of all three TET enzymes. In the developmentally relevant GN11 cell line versus adult GT1-7 neurons, we found that Tet mRNA continues to increase into adulthood (Linscott &

Chung, 2019). We next measured the ability of TET proteins to demethylate methylated cytosine residues on a genome-wide scale. We therefore performed dot-blot assays in genomic DNA samples from either the embryonic E10.5 versus E13.5 OP. While we found no change in 5mC pixel intensity, there was a significant increase in 5hmC in the E13.5 OP, indicating that 5hmC is critical for OP maturation (Linscott & Chung, 2019).

These results parallel studies in GN11 and the GT1-7 GnRH-secreting cell lines performed in 2016, which found higher expression levels of only Tet2 mRNA, albeit, Tet1 mRNA was not significant at p = 0.09 (Kurian et al., 2016). These results may be conflicting due to differences in passage number however, we believe that our cell line expression data parallels the OP more closely (Linscott & Chung, 2019). However, both studies require a more detailed investigation of specific genes that are demethylated during GnRH neuron onset. Specifically,

Kurian showed that in adults, GnRH neuron-specific Tet2 knockout animals reduced male

167 fecundity, however, no alterations in pubertal timing of LH in either sex was observed (Kurian et al., 2016). Together, these results suggest that all three TETs likely act on GnRH neurons during development and in adulthood. TET target genes should therefore be systematically assessed for both binding events, alterations in histone modifications, and in RNA processing, as TETs have been shown to regulate these diverse functions (Basanta-Sanchez et al., 2017; M. A. Hahn et al.,

2013; Miao et al., 2016; Shen et al., 2014; Wu et al., 2011).

Future Directions

Collectively, our studies demonstrate the intricacies of Fgf8 transcription during a critical window of activation which is required for GnRH neuron onset. Specifically, we show that Fgf8 is not activated by previously identified transcription factors as identified in many other systems,

(i.e., prostatic, breast cancer cells, embryonic trunk) but by discrete and timely chromatin modifications. We showed that Fgf8 is bivalent during GnRH neuron emergence in vivo, which is controlled by TET1. Moreover, TETs drove a demethylation pattern on the Fgf8 promoter at both the transcription start site and at a distal region. However, the function of the distally demethylated promoter region remains to be elucidated. For example, are there accessory regulatory regions near the distal site, and moreover, are they required for Fgf8 activation? In a broader context, timing of Fgf8 transcription should also be evaluated in reproductive output, number of GnRH neurons, and the role of Fgf8 transcription in both developing and adult GnRH neurons.

Although studies in the OP were not GnRH neuron specific, we determined that Fgf8 was bound by TET1 at E9.5 and at E13.5, and that the promoter region of Fgf8 was demethylated with embryonic age. In support, we found that TET1-3 mRNA was upregulated in the developing

168

OP, which additionally underwent a genome-wide demethylation event. Future studies should perform these studies in developing GnRH neurons specifically. For example, using single cell- sequencing coupled with single cell ChIP-seq will more accurately determine the expression and chromatin state of the developing GnRH neuronal population. These results could further be translated during pubertal stages to target the GnRH neuronal population epigenetically, with the use of direct genome editing.

Currently, it is not known whether Fgf8 is expressed in GnRH neurons While studies in the early OP have shown Fgf8 expressing cells in the respiratory epithelium (Bachler &

Neubuser, 2001), no Fgf8 mRNA is expressed in the olfactory epithelium where early GnRH neurons can be detected. These early studies indicate that FGF8 signaling likely diffuses from the respiratory epithelium to the olfactory epithelium to promote the emergence of GnRH neurons. Indeed, studies in the OP demonstrated the presence of FGFR1 on the cell membrane of

GnRH expressing neurons (John C. Gill et al., 2004). Moreover, in human pluripotent stem cells,

FGF8 treatment is necessary for GnRH neuron programming (Lund et al., 2016), indicating that

FGF8 plays a role in fate specification of GnRH neurons in-vitro. Indeed, unpublished studies from our laboratory demonstrate that FGF8b treatment of migratory GN11 neurons induces a

MAPK pathway. Therefore, we conclude that Fgf8 acts as a distal morphogen which likely regulates GnRH cell fate and is secreted and/or produced from nearby cells, which then acts on the FGFR1 receptor located on the GnRH neuron.

While previous studies have elucidated the role of TET2 in GnRH neurons, a conditional triple knock out model in GnRH neurons should be utilized during GnRH emergence and pubertal onset to determine the role of hydroxymethylation in GnRH neuron function. These studies can more accurately predict target genes, and/or RNA targets that harbor 5hmC

169 modifications, which have recently been shown to be catalyzed by the TET family (Basanta-

Sanchez et al., 2017). Overall, understanding the role of the epigenome during critical periods of development such as GnRH neuron onset and puberty, will further elucidate environmental factors or pharmaceuticals that may impact the epigenome.

Conclusions

GnRH neuron development as well as pubertal initiation are two major developmental stages that are coordinated in a precise fashion. Epigenetics remains a central avenue of research to explain the mechanisms behind these complex processes. Our recent research indicate that transient OP

Fgf8 transcription is due to the coordinated actions of DNMT3b, TET1 and EZH2, as well as bivalent histone modifications. Timely epigenetic modifications to the Fgf8 locus proved to be critical in both transcriptional activation and repression, which coincided with the emergence of

GnRH neurons in the OP. Interestingly, the GnRH promoter undergoes similar demethylation event during both GnRH neuron ontogenesis and again during pubertal onset. We conclude that the epigenome is a major molecular regulator in GnRH neuron development and function, which could affect the activities of the GnRH neuron system as it pertains to the HPG-axis and reproductive success. The evidence also points towards a common epigenetic “timing” mechanism that precisely controls reproductively relevant gene transcription; which begs further epigenetic investigations of not only development and pubertal onset, but timing of LH surges, hormonal feedback from gonads, and input from KP neuron populations onto GnRH neurons.

170

REFERENCES

Albuisson, J., Pecheux, C., Carel, J. C., Lacombe, D., Leheup, B., Lapuzina, P., Bouchard, P.,

Legius, E., Matthijs, G., Wasniewska, M., Delpech, M., Young, J., Hardelin, J. P., &

Dode, C. (2005). Kallmann syndrome: 14 novel mutations in KAL1 and FGFR1 (KAL2).

Hum Mutat, 25(1), 98-99.

Alexandre, P., Bachy, I., Marcou, M., & Wassef, M. (2006). Positive and negative regulations by

FGF8 contribute to midbrain roof plate developmental plasticity. Development, 133(15),

2905-2913. doi: 10.1242/dev.02460

Ang, H. L., Deltour, L., Žgombić‐Knight, M., Wagner, M. A., & Duester, G. (1996). Expression

patterns of class I and class IV alcohol dehydrogenase genes in developing epithelia

suggest a role for alcohol dehydrogenase in local retinoic acid synthesis. Alcoholism:

Clinical and Experimental Research, 20(6), 1050-1064.

Antunes, C., Sousa, N., Pinto, L., & Marques, C. J. (2019). TET enzymes in neurophysiology

and brain function. Neurosci Biobehav Rev, 102, 337-344. doi:

10.1016/j.neubiorev.2019.05.006

Aylwin, C. F., Toro, C. A., Shirtcliff, E., & Lomniczi, A. (2019). Emerging Genetic and

Epigenetic Mechanisms Underlying Pubertal Maturation in Adolescence. J Res Adolesc,

29(1), 54-79. doi: 10.1111/jora.12385

Bachler, M., & Neubuser, A. (2001). Expression of members of the Fgf family and their

receptors during midfacial development. Mech Dev, 100(2), 313-316.

Basanta-Sanchez, M., Wang, R., Liu, Z., Ye, X., Li, M., Shi, X., Agris, P. F., Zhou, Y., Huang,

Y., & Sheng, J. (2017). TET1-Mediated Oxidation of 5-Formylcytosine (5fC) to 5-

171

Carboxycytosine (5caC) in RNA. Chembiochem, 18(1), 72-76. doi:

10.1002/cbic.201600328

Beermann, F., Kaloulis, K., Hofmann, D., Murisier, F., Bucher, P., & Trumpp, A. (2006).

Identification of evolutionarily conserved regulatory elements in the mouse Fgf8 locus.

Genesis, 44(1), 1-6. doi: 10.1002/gene.20177

Belsham, D. D., Evangelou, A., Roy, D., Duc, V. L., & Brown, T. J. (1998). Regulation of

gonadotropin-releasing hormone (GnRH) gene expression by 5alpha-dihydrotestosterone

in GnRH-secreting GT1-7 hypothalamic neurons. Endocrinology, 139(3), 1108-1114.

doi: 10.1210/endo.139.3.5846

Bernstein, B. E., Mikkelsen, T. S., Xie, X., Kamal, M., Huebert, D. J., Cuff, J., Fry, B., Meissner,

A., Wernig, M., Plath, K., Jaenisch, R., Wagschal, A., Feil, R., Schreiber, S. L., &

Lander, E. S. (2006). A bivalent chromatin structure marks key developmental genes in

embryonic stem cells. Cell, 125(2), 315-326. doi: 10.1016/j.cell.2006.02.041

Bhattacharyya, S., & Bronner-Fraser, M. (2008). Competence, specification and commitment to

an olfactory placode fate. Development, 135(24), 4165-4177. doi: 10.1242/dev.026633

Bird, A. (2007). Perceptions of epigenetics. Nature, 447(7143), 396-398.

Bird, A. P., & Wolffe, A. P. (1999). Methylation-induced repression--belts, braces, and

chromatin. Cell, 99(5), 451-454.

Boehm, U., Bouloux, P. M., Dattani, M. T., de Roux, N., Dode, C., Dunkel, L., Dwyer, A. A.,

Giacobini, P., Hardelin, J. P., Juul, A., Maghnie, M., Pitteloud, N., Prevot, V., Raivio, T.,

Tena-Sempere, M., Quinton, R., & Young, J. (2015). Expert consensus document:

European Consensus Statement on congenital hypogonadotropic hypogonadism--

172

pathogenesis, diagnosis and treatment. Nat Rev Endocrinol, 11(9), 547-564. doi:

10.1038/nrendo.2015.112