Báo cáo khoa học: Gonadotropin-releasing hormone: regulation of the GnRH gene pot

Abbreviations AP-1, activator protein-1; AR, androgen receptor; atRA, all-trans-retinoic acid; C ⁄ EBP, CCAAT ⁄ enhancer-binding protein; CREB, cAMP response element-binding protein; DHE

Gonadotropin-releasing hormone: regulation of the

GnRH gene

Vien H Y Lee, Leo T O Lee and Billy K C Chow

School of Biological Sciences, The University of Hong Kong, China

Introduction

Gonadotropin-releasing hormone (GnRH) is a central

regulator in the hypothalamic–pituitary–gonadal axis

of the reproductive hormonal cascade It is expressed

in a discrete population of neurosecretory cells located

throughout the basal hypothalamus of the brain, and

is released into the hypothalamo-hypophyseal portal

circulation in a pulsatile manner and in surges duringthe female preovulatory period [1] The releasedGnRH is transported to the anterior pituitary gland,where the hormone binds to its receptor on thegonadotropes This triggers the synthesis and release

of the gonadotropins luteinizing hormone (LH) andfollicle-stimulating hormone (FSH), which are respon-sible for gonadal steroidogenesis and gametogenesis

Keywords

estrogen; follicle-stimulating hormone;

GnRH; gonadotropin; luteinizing hormone;

PKC signalling; progesterone; promoter;

steroid hormone; transcriptional regulation

Correspondence

B K C Chow, School of Biological

Sciences, The University of Hong Kong,

Pokfulam Road, Hong Kong, China

by gonadotropins and gonadal sex steroids Other physiological stimuli,e.g insulin and melatonin, will also be discussed

Abbreviations

AP-1, activator protein-1; AR, androgen receptor; atRA, all-trans-retinoic acid; C ⁄ EBP, CCAAT ⁄ enhancer-binding protein; CREB, cAMP response element-binding protein; DHEA, dehydroepiandrosterone; DHT, dihydrotestosterone; Dlx2, distal-less homeobox 2; DREAM, downstream regulatory element antagonist modulator; E2, 17b-estradiol; EMSA, electrophoretic mobility shift assay; ER, estrogen receptor; FSH, follicle-stimulating hormone; GABA, c-aminobutyric acid; GnRH, gonadotropin-releasing hormone; GRG, Groucho-related gene; hCG, human chorionic gonadotropin; hGLC, human granulosa-luteal cell; hGnRH-I, human gonadotropin-releasing hormone type I; IGF-I, insulin-like growth factor-I; LH, luteinizing hormone; mGnRH-I, mouse gonadotropin-releasing hormone type I; Msx, muscle segment homeobox; NIRKO, neuron-specific insulin receptor knockout; NMDA, N-methyl- D -aspartic acid; NO, nitric oxide; nPRE, negative progesterone response element; Oct-1, octamer-binding transcription factor-1; Otx, orthodenticle homeobox; P4, progesterone; POU, homeodomain protein family of which the founder members are Pit-1, Oct-1/2 and Unc-86 ; PKA, protein kinase A; PKC, protein kinase C; POA, preoptic area; PR,

progesterone receptor; RA, retinoic acid; RAR, retinoic acid receptor; RARE, retinoic acid response element; rGnRH-I, rat releasing hormone type I; RXR, retinoid X receptor; TPA, 12-O-tetradecanoyl phorbol-13-acetate; b-gal, b-galactosidase.

gonadotropin-During embryogenesis, GnRH-expressing neurons

arise in the olfactory placode, migrate into the preoptic

area (POA), and then extend axons to the median

eminence [2] The hypothalamic expression of GnRH

increases gradually during postnatal development and

puberty, and is believed to be crucial for the onset of

puberty [3]

GnRH is a peptide hormone composed of 10 amino

acids

(pGlu-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-Gly-NH2) Gene expression first gives rise to the

prepro-GnRH polypeptide, which consists of a signal peptide,

a functional decapeptide, an amidation⁄ proteolytic

processing signal (Gly-Lys-Arg), and a

GnRH-associ-ated peptide [4] According to the differences in amino

acid sequences, localizations and embryonic origins, 24

GnRHs have been identified in the nervous tissues,

from vertebrates to protochordates [5,6] Despite the

above divergences, all of these variants are

decapep-tides that share highly similar structures Generally,

two or three forms of these GnRHs can be found in

most vertebrate species

It had been thought that mammals have only one,

classical, form of GnRH (GnRH-I) GnRH-I molecules

in different mammals have identical amino acid

sequences, except in guinea pig, in which the second

and seventh amino acids are substituted [7] In humans,

this gene is located on chromosome 8p11.2-p21, with

four exons that contain a 276 bp ORF coding for a

precursor protein of 92 amino acids Recently,

how-ever, a second form of GnRH (GnRH-II) was

identi-fied As GnRH-II was originally isolated from chicken

brain, it was termed chicken GnRH-II [8] GnRH-II is

encoded by a different gene and differs from GnRH-I

by amino acids [5,7–9] GnRH-II is the most

ubiqui-tous peptide of the GnRH neuropeptide family, being

present in animals from jaw fish to humans The highly

conserved amino acid sequence of GnRH-II in a wide

range of species and over millions of years of evolution

suggests the importance of this neuropeptide The

tis-sue distribution pattern of GnRH-II is dissimilar to

that of GnRH-I Whereas GnRH-I is expressed mostly

in the brain, the expression level of GnRH-II is much

higher in other organs

Regulation of GnRH gene expression

In view of the fact that GnRH is essential for

repro-ductive processes, understanding the control of its

synthesis and release is therefore of the utmost

impor-tance However, it is difficult to study the regulation

of GnRH gene transcription in vivo, due to the

scar-city and scattered distribution of the GnRH neurons

In the past, immortalized GnRH-expressing neuronal

cell lines have probably been the only effective andmanageable resources with which to explore mecha-nisms regulating the expression, synthesis and release

of GnRH Studies on transcriptional regulation of theGnRH gene have been performed largely in GnRH-secreting cell lines, such as GT1, GT1-7, and humangranulosa-luteal cells (hGLCs) [1,10–14] The GT1 cell

is recognized as a good model for studying specific expression of the GnRH gene, as GT1 cellsretain many characteristics of in vivo GnRH neurons.These include distinct neuronal morphology [15],expression of differentiated neuronal markers [16], thepulsatile release of GnRH in cell culture [17,18], andsecretion of GnRH in response to particular signals

neuron-It should be noted that although in vitro studies incell lines have been widely employed, they are unlikely

to resemble the actual complexity of gene regulation

in the brain or other organs Only recently has thedevelopment of various transgenic mice enabled theinvestigation of GnRH gene expression and regulation

in vivo [19–21]

In this minireview, we summarize studies regardingGnRH-I and GnRH-II gene regulation, including essen-tial cis-acting elements in the promoter, and also theinteraction of transcription factors in achieving thebasal expression levels In addition, other components

of the hypothalamic–pituitary–gonadal axis with roles

in the control of GnRH-I gene expression will also bediscussed These include gonadotropin, gonadal sexsteroids and other physiological regulators

GnRH-I

Analysis of the promoters of the GnRH-I and GnRH-IIgenes showed that they contain essentially differentputative transcription factor-binding sites that areimportant for their basal transcription activities, sug-gesting that the two genes are probably differentiallyregulated Because of the recent discovery of GnRH-II

as a new isoform of GnRH, the majority of the studieshave been done on the GnRH-I promoter, and thereare only a few regarding regulation of the GnRH-IIgene

The GnRH-I promoterAmong the studies on the transcriptional regulation ofGnRH genes, most have been performed on GnRH-I.The 5¢-flanking region of the GnRH-I gene is highlyhomologous between species [22], especially human,rat and mouse A summary of the essential elements inthe promoter regions of the rat, mouse and humanGnRH-Igenes is given in this minireview (Figs 1–3)

+273 –5500 –2100 –1700

Fig 2 Diagrammatic representation of the mGnRH-I gene 5¢-region The promoter region is indicated by the green box [20] and the enhancer by the blue box [19] Also, the locations of key regulatory elements and their functional significance are listed.

The transcription start sites

For the determination of the transcription initiation

site in the rat and mouse GnRH genes, primer

exten-sion analysis was employed [23] In the study, the first

exons in the rat and mouse GnRH genes were found to

be 145 and 58 bp respectively, using polyA+ RNA

from rat hypothalamus and total RNA from mouse

hypothalamic GT1-7 cells In humans, studies showed

that transcription of the GnRH gene can be initiated at

two distinct transcription start sites in the

hypothala-mus and nonhypothalamic tissues such as ovary, testes,

placenta and mammary gland [24–26] In the

hypothal-amus, the transcription start site was characterized at

61 bp upstream of the first exon–intron junction,

whereas a discrete upstream transcription start site,

which is 579 bp upstream from the hypothalamic start

site, was identified later in a human placental tumor

cell line (JEG) and a human breast tumor cell line

(MDA), using primer extension and RT-PCR assays

The placental GnRH cDNAs were reported to have a

longer 5¢-UTR than that found in the hypothalamus

Radovick et al found that the alternative mRNA was

produced by differential splicing of the GnRH gene

The first intron is removed in the hypothalamus,

whereas it is retained in the placenta Also, the human

upstream promoter has been found to have a higher

level of transcriptional activity than the downstream

one However, it is not homologous to the upstream

region of rat and mouse genes, and there is no

evi-dence showing the use of the upstream promoter in

any of the rat and mouse tissues [24] A later study

was carried out to compare the GnRH gene of

non-human primates with that of non-humans [25] The study

showed the presence of an upstream transcription start

site in the cynomolgus monkey, 504 bp upstream of

the hypothalamic promoter, and 75 bp downstream of

the human upstream start site Sequence analysis

showed that the cynomolgus monkey and the human

upstream promoter share high similarity (94%)

Rat GnRH-I promoter

Two key regions, including a proximal promoter and a

distal enhancer, have been identified in the rat GnRH-I

(rGnRH-I) gene that are important for gene

transcrip-tion (Fig 1) The promoter is located 173 bp upstream

of the transcription start site [22] It is evolutionarily

conserved, with about 80% nucleotide homology

among human, rat and mouse The 300 bp enhancer is

located at)1863 to )1571 bp relative to the

transcrip-tion start site It provides 50–100-fold activatranscrip-tion of

GnRHgene transcription as compared to the activity of

the promoter alone [22,27] The rat promoter has beenwidely studied for many years; specific binding sites for

a number of different transcription factors have beenfound within the promoter and enhancer

Mouse GnRH-I promoterThe two promoter regions of GnRH-I are highly con-served in rat and mouse (Fig 2) The development oftransgenic mice provided in vivo models for GnRH generegulation studies Transgenic mice carrying variousdeletion fragments of the mouse GnRH-I (mGnRH-I)gene fused to a reporter gene have been used foridentifying essential sequences for GnRH-I gene expres-sion in GnRH neurons and in the ovary in vivo [19,20].Pape et al demonstrated that a 5.5-kb fragment of the5¢-region of the mGnRH-I gene was sufficient to targetb-galactosidase (b-gal) and thus GnRH-I expression inabout 85% of GnRH neurons [19] Deletion of the5¢-flanking sequence to 2.1 kb resulted in a 40%reduction in the number of b-gal-expressing GnRHneurons This suggests that enhancer element(s) arepresent in the region between )5.5 and )2.1 kb of themGnRH-I gene More importantly, further 5¢-deletion

to)1.7 kb resulted in total loss of b-gal detection Thisindicates that the 400 bp region ()2.1 to )1.7 kb) is acritical enhancer region for the mGnRH-I gene in mousebrain in vivo Later, Kim et al also worked on trans-genic mice, and demonstrated that specific expression ofthe mGnRH-I gene in the hypothalamus and ovarydepends on a proximal region ()1005 bp) of themGnRH-I promoter in the hypothalamus and ovary[20] This indicates the presence of elements specific tothe hypothalamus and ovary within the)1005 bp region

of the mouse promoter Moreover, through generation

of transgenic mice with deletion fragments, the regionbetween)3446 and )2078 bp, which was found to haveabout 90% homology with the rGnRH enhancer, wasshown to be an enhancer for in vivo expression of hypo-thalamic mGnRH, as well as a repressor that repressesmGnRHgene expression in the mouse ovary

Human GnRH-I promoterSequence alignment has revealed similarities and 8 dif-ferences between the human GnRH-I (hGnRH-I) geneand the rGnRH-I gene [28] (Fig 3) Within the distalpromoter of the hGnRH gene, three regions ()3036 to)2923 bp, )2766 to )2539 bp, and )1775 to )1552 bp)were found to be similar to sequences in the rGnRHpromoter In the proximal promoter region, the )343

to +8 bp region of the hGnRH-I gene was found tohave marked homology with the )332 to +96 bp

region of the rGnRH-I gene However, the sequence

between )1552 and )579 bp in the hGnRH-I gene has

little similarity with the rat promoter In Gn10 cells,

Dong et al identified the hGnRH promoter in the)551

to +1 bp region [24] In agreement with this, Kepa

et al found that the )3832 to +8 bp fragment of the

hGnRH-I gene gave high levels of expression

of reporter and GnRH-I genes in GT1-7 cells [28]

A 5¢-deletion to )1131 bp had no effect, whereas a

3¢-deletion to )350 bp led to a significant reduction

(70%) of promoter activity, showing the importance of

the )1131 to )350 bp region for regulation of the

hGnRH-I gene In in vivo studies using transgenic mice

with 5¢-deletion fragments of the hGnRH-I gene, the

)1131 to )484 bp region was found to include

cell-spe-cific elements for hGnRH gene expression [29] Later

studies further determined that the )992 to )763 bp

region is essential and sufficient for specific hGnRH

gene expression in GnRH neurons [21]

Regulation of GnRH-I by essential transcription

factors

Oct-1, Msx1, Dlx2 and cofactors

Octamer-binding transcription factor-1 (Oct-1) plays a

critical role in the regulation of rGnRH-I transcription,

binding functional elements in the proximal promoter

region [12] DNase I protection experiments revealed

that a 51-bp sequence ()76 to )26 bp) conferred a

20-fold induction of the rGnRH-I gene in GT1-7 cells This

region contains an octamer-like motif ()47 to )40 bp)

to which Oct-1, a member of the homeodomain protein

family of which the founder members are Pit-1, Oct-1/2

and Unc-86 (POU), was found to bind Oct-1 was also

found to bind the octamer motifs in another promoter

region ()110 to )88 bp) Within the enhancer of the

rGnRH-I gene, two POU homeoprotein Oct-1-binding

sites, OCT1BS-a ()1785 to )1771 bp) and OCTBS-1b

()1702 to )1695 bp), which share a 6-bp sequence with

the octamer consensus sequence (ATGCAAAT), were

identified [30] Electrophoretic mobility shift assays

(EMSAs) showed the binding of Oct-1 proteins to both

sites Block mutation of OCT1BS-a resulted in a

signifi-cant reduction (95% reduction) in transcriptional

activ-ity, showing that OCT1BS-a is the most crucial element

for transcriptional activity of the GnRH-I gene

enhan-cer However, mutation of OCT1BS-b had no effect on

the enhancer activity Consistently, OCT1BS-b was also

reported to be not involved in basal or unstimulated

enhancer activity of the rGnRH-I gene in GT1-7 cells

[31] Mutation of OCT1BS-b resulted in elimination of

repression by the glutamine–NO–cGMP signaling

path-way, but did not influence the nonrepressed GnRH gene

expression This suggested that OCT1BS-b may play arole in modulated but not basal transcriptional activity

of the rGnRH-I gene

Two other homeodomain proteins, Mex1 and less homeobox 2 (Dlx2), have also been identified asbeing responsible for GnRH-I gene regulation Withinthe proximal promoter and the enhancer of therGnRH-I gene, four conserved consensus homeo-domain sites (ATTA) ()41 to )38 bp, )54 to )51 bp,)1620 to )1617 bp, and )1634 to )1631 bp) have beenidentified as being essential for basal and cell-specificexpression of rGnRH-I in GT1-7 cells [32,33] Also,Givens et al found that muscle segment homeobox(Msx) and Dlx, which are members of the antennape-dia class of non-Hox homeodomain transcriptionfactors, bind to the ATTA consensus sequence [34].Msx1 is found as a repressor, whereas Dlx2 is an acti-vator, and they functionally antagonize each other bycompeting for the ATTA elements in the rGnRH-Igene regulatory regions

distal-The majority of the identified transcriptional tors of the GnRH genes are homeodomain proteinswith promiscuous DNA-binding properties, and mostare not solely expressed in GnRH neurons To achievespecific activity of the promoter and target expression

regula-of GnRH in GnRH neurons, specific interactions regula-ofthe transcription regulators with specific cofactors arerequired [35] These cofactors can enhance or inhibitthe interactions between the homeodomain proteinsand the transcriptional regulatory regions of the GnRHpromoter and⁄ or enhancer to achieve specific GnRHexpression in the hypothalamic GnRH neurons and innonhypothalamic tissues, such as the ovary

Ravel-Harel et al reported that the Groucho-relatedgene (GRG) proteins, which belong to the GRG fam-ily of coregulators, associate with the GnRH promoter

in vivo and interact with Oct-1 and Msx1 in GT1-7cells [36] GRG proteins mediate the dynamic switchbetween activation and repression of GnRH-I tran-scription Using glutathione S-transferase pull-downassays, the long-form GRG proteins (GRG1 andGRG4) and the short truncated form (GRG5) werefound to interact with Oct-1 and Msx1, probablythrough the POU domain of Oct-1 and the engrailedhomology domain of Msx1 As shown in overexpres-sion studies, the long GRG forms are repressors ofGnRH-I gene transcription They repress the Oct-1-mediated activation and act as corepressors of Msx1,mediating downregulation of GnRH-I expression Incontrast, the short form GRG5 is an enhancer thatreverses the repressive activity of GRG4

A three amino acid loop extension (TALE) domain transcription factor, Pbx1b, was also found to

homeo-be a cofactor of Oct-1 by using a yeast two-hybrid

sys-tem [37] Moreover, Pbx1b contains the Meis

homolo-gous regions in the N-terminus and the Prep1

homeodomain in the C-terminus for interaction with

its cofactors Prep1 and Meis [38] A GST pull-down

assay showed the in vitro interaction of Oct-1 with

Pbx1b and its cofactor Prep1 EMSA showed

hetero-dimers containing TALE proteins, Pbx⁄ Prep1 and

Pbx⁄ Meis1, in GT1-7 nuclear extract bound to four

binding sites within both the promoter (at )100 and

)75 bp) and enhancer (at )1749 and )1603 bp) These

binding sites are in close proximity to or even overlap

with the Oct-1 sites Both Pbx1 and Prep1 are

coacti-vators of Oct-1 in GnRH-I expression, because

coex-pression of Oct-1 with either Pbx1b or Prep1 resulted

in significant activation of GnRH-I gene transcription,

whereas no significant change was observed when these

constructs were overexpressed individually without

Oct-1 Within those Pbx1⁄ Prep1-binding motifs, only

mutation in the )1749 bp binding site can eliminate

the activation, indicating that the transactivation is

specifically dependent on this motif

Otx2

In the proximal promoter of the rGnRH-I gene, an

orthodenticle homeobox (Otx)⁄ bicoid site ()153 to

)146 bp) is conserved across several vertebrate species

[14] This element in the rGnRH-I promoter was found

to bind Otx2 proteins The promoter activity was

sig-nificantly reduced in both the promoter alone and the

promoter with enhancer constructs after the Otx⁄

bicoid site was mutated Moreover, overexpression of

Otx2 in GT1-7 cells resulted in induction of rGnRH-I

promoter activity This showed that the Otx⁄ bicoid

element was important for basal and also

enhancer-dri-ven transcription of the GnRH-I gene Recently, the

critical role of Otx2 in regulating tissue-specific

expres-sion of the mGnRH-I gene has been discovered In a

transgenic mice study, high luciferase activities were

only detected in the hypothalamus and gonads when a

DNA construct containing the)356 to +28 bp region

of the mGnRH-I gene fused to a luciferase reporter

gene was used to generate transgenic mice [39]

How-ever, transgenic mice with the 5¢-deletion construct

)249 to +28 bp showed high luciferase expression

only in gonads This difference indicated that the

DNA sequence between )356 and )249 bp was

essen-tial for neuron-specific expression of the GnRH-I gene

Within this region, Kim et al identified two consensus

Otx2-binding sites, a low-affinity binding site (TTATC,

)319 to )315 bp) and a high-affinity binding site (TA

ATCC, )257 to )252 bp) EMSA demonstrated that

Otx2 binds both consensus sites specifically expression of Otx2 in GN11 cells increased mGnRHgene transcriptional activity by more than fivefold.Moreover, in vivo studies using Otx2-binding sites inmutated transgenic mice showed that elimination ofthese Otx2 sites resulted in reduced GnRH promoteractivities in the mouse brain This further confirmedthe importance of Otx2 binding for appropriateneuronal expression of GnRH

Over-Brn2Another POU homeodomain protein, Brn2, expressed

in the hypothalamus and olfactory tissues, has alsobeen found to regulate hGnRH-I gene expression [21]

In vivo studies of transgenic mice showed the regionbetween )992 and )795 bp to be important for GnRHneuron-specific expression of the hGnRH-I gene.Within this region, two POU protein-binding sites()925 to )916 bp and )867 to )858 bp) have beenidentified These sites have high homology with theBrn2 consensus binding site EMSA showed that Brn2proteins in NLT nuclear extract bound to the Brn2consensus binding site, but not to a mutated Brn2 con-sensus site Also, overexpression of Brn2 increasedmGnRHmRNA expression in cultured GnRH neuronsand GN11 cells, and also enhanced hGnRH promoteractivities in GN11 cells

Fos and CREB

A previous study demonstrated that treatment ofGT1-7 neurons with human chorionic gonadotropin(hCG), a GnRH inhibitor, resulted in an increase inphosphorylated Fos, Jun and cAMP response element-binding protein (CREB) [40] Overexpression studiesshowed that Fos and CREB, but not Jun, inhibitedrGnRH-I promoter activity in a dose-dependent man-ner in the)3026 to +116 bp construct, and the inhibi-tory action of CREB was more effective than that ofFos [41] However, these proteins were found not tobind to the hCG responsive region ()126 to )73 bp)

of the rGnRH-I proximal promoter, as shown bysupershift assays using antibodies against these pro-teins This suggests that Fos and CREB might bind toother motifs or might interact with other proteins thatbind the rGnRH-I promoter to achieve its regulation

GATA-4, GBF-A1/A2 and GBF-B1

In GT1 cells, two GATA factor-binding motifs thatoccur in tandem repeats have been found within therGnRH-I enhancer region (GATA-A, )1710 to

)1715 bp; GATA-B, )1743 to )1748 bp) [10]

Muta-tion analysis demonstrated that both GATA sites are

functionally important and that one factor, GATA-4,

present in GT1 cells, can interact with the

GATA-binding motifs Later, Lawson et al also reported the

binding of two GATA factors, GBF-A1⁄ A2 and

GBF-B1, to the GATA factor-binding sites in the

GnRH-I enhancer [42] GATA-4 and GBF-B1 were

found to be necessary for full enhancer activity; both

factors are able to activate the GnRH-I promoter

However, GBF-B1 was also shown to modulate the

GATA-4-mediated activation of the GnRH-I enhancer

by competing with GATA-4 for binding to the

GATA-binding motifs

Protein kinase C (PKC) signaling pathway

The above studies revealed multiple transcription

fac-tors binding to a number of regulatory elements in the

proximal promoter and the enhancer that are crucial

for the regulation of GnRH-I gene transcription Other

studies have shown that GnRH-I gene regulation is

also mediated through the PKC signaling pathway

(Fig 4) Treatment of GT1-7 cells with the protein

kinase A (PKA) pathway activator forskolin (10 lm)

did not produce any effect on the rGnRH-I mRNA

levels, whereas treatment with the PKC pathway

acti-vator 12-O-tetradecanoyl phorbol-13-acetate (TPA)

(100 nm) caused a significant reduction (70%) in

rGnRH-ImRNA expression [43] Bruder et al showed

that TPA caused dose- and time-dependent repression

of rGnRH-I promoter activity and a decrease in

rGnRH-I transcript levels, which was mediated byincreased c-fos and c-jun mRNA levels [44] This TPA-mediated repression was found to be dependent on theproximal promoter region at )126 to )73 bp, withinwhich an activator protein-1 (AP-1) site is present (at)99 bp) However, Fos and Jun have been shown not

to bind the AP-1 site directly, but to interact withother protein(s) that bind to this site in the proximalpromoter Recently, not only the proximal promoter,but also the enhancer, of the rGnRH-I gene was found

to participate in PKC repression [45] Various elements within the enhancer region, as describedabove, are required for the repression of rGnRH-Iexpression by PKC These include Oct-1, Prep⁄ Pbx1a,and Dlx2 TPA causes activation of PKC, which inturn leads to increased phosphorylation of these tran-scription factors, and therefore reduces binding to theirinteracting sites within the enhancer region of therGnRH-I gene The study also revealed a novel site()1793 to )1785 bp), to which an unknown proteinfrom GT1-7 nuclear extract bound, was involved inthis PKC repression of the rGnRH-I gene expression

cis-Pulsatile GnRH-I gene expressionGnRH-I is released in a pulsatile manner in the hypo-thalamus This intrinsic property of GnRH neuronswas first observed in isolated hypothalamic fragments[46,47] and dispersed hypothalamic neuron cultures[48] Pulsatility is not restricted only to GnRH-Irelease, but is also associated with GnRH-I geneexpression Recent studies revealed that GnRH-I

GnRH-I

Fig 4 PKC repression of rGnRH-I gene expression The PKC responsive sites within the proximal promoter and the enhancer of the rGnRH-I gene are indicated by colored boxes.

promoter activity operates in a pulsatile manner

[49,50] In fact, the secretory pulse of GnRH-I and the

episodic GnRH-I gene expression were found to be

clo-sely associated

Analysis of the rGnRH-I promoter revealed that a

certain region between )2012 and )1597 bp, which

includes the enhancer, was responsible for the

pulsatili-ty [28] Within this region, three Oct-1-binding sites

were identified [30,51] OCT1BS-a ()1785 to )1771 bp)

and OCTBS-1b ()1702 to )1695 bp) were shown by

Clark and Mellon to bind the Oct-1⁄ POU

homeo-protein More recently, a new site, OCT1BS-c ()1569

to)1562 bp), was also found to bind Oct-1 [51]

How-ever, only OCT1BS-a and OCT1BS-c, and not

OCT1BS-b, were shown to be necessary for the

pulsa-tility in mutation analysis

A recent study also demonstrated the involvement

of Ca2+ and a novel Ca2+-binding protein,

down-stream regulatory element antagonist modulator

(DREAM), in GnRH-I pulsatility [52] DREAM was

identified as being responsible for the GnRH-I

pulsa-tility, as it was shown to be part of the OCT1BS-b

binding complex, an essential element in the GnRH

enhancer for promoter pulse Also,

immunoneutraliza-tion of DREAM in single GT1-7 cells resulted in a loss

of episodic GnRH-I gene expression An L-type Ca2+

blocker, nimodipine, which markedly reduced the

GnRH-I secretory pulse, was also shown to abolish

GnRH-I gene expression pulses These findings

sug-gested that DREAM, via Ca2+, may serve as a basis

for the communication between cytoplasm and nucleus

that links the pulsatile secretion and pulsatile

expres-sion of GnRH-I

GnRH-II

The GnRH-II promoter

Cheng et al found the core promoter region of the

human GnRH-II gene to be located between )1124

and )750 bp relative to the translation start codon bytransient transfection studies in neuronal medulloblas-toma TE-671 cells, placental choriocarcinoma JEG-3cells and ovarian carcinoma OVCAR-3 cells [53](Fig 5) Moreover, the untranslated exon 1 ()793 to)750 bp) was found to be an enhancer element forstimulation of GnRH-II gene expression [53]

Regulation of GnRH-II by essential transcriptionfactors

AP-1 and AP-4Within the untranslated exon 1 of the hGnRH-II gene,two E-box-binding sites ()790 to )785 bp and )762 to)757 bp) and one Ets-like element ()779 to )776 bp)were found [53] These three regulatory elements work

in a cooperative manner for basal hGnRH-II gene scription Studies showed in vitro specific binding ofthe basic helix–loop–helix transcription factor AP-1 tothe two E-box-binding sites, whereas an unknownprotein bound to the Est-like element EMSA usingTE-671 nuclear extracts with oligonucleotides contain-ing the two E-box motifs showed the formation ofDNA–protein complexes, which was abolished by aconsensus AP-4-binding sequence Also, in vitro trans-lated human AP-4 proteins bound to the two E-box-binding sites formed a complex with similarelectrophoretic mobility to that formed with TE-671extracts Overexpression studies revealed that AP-4 is anenhancer that upregulates hGnRH-II promoter activity

tran-p65, retinoic acid receptor-a (RARa) and retinoid Xreceptor-a (RXRa)

Recently, our group has identified a repressor elementGII-Sil within the first introns ()641 to )636 bp) of thehGnRH-II promoter [54] EMSA showed that proteins

in TE-671 nuclear extracts formed two specific DNA–protein complexes with GII-Sil In vitro supershiftassays and an in vivo chromatin immunoprecipitation

+1 –750

Fig 5 Diagrammatic representation of the

hGnRH-II gene 5¢-region The promoter

region is indicated by the green box [51]

and the enhancer by the yellow box [51].

Also, the locations of key regulatory

ele-ments and their functional significance are

listed.

assay also showed that nuclear factor kappa B (NF-jB)

p65 subunit, and retinoic acid receptors (RARa and

RXRa), bind to the GII-Sil element Also, functional

analysis revealed that p65 is a downregulator of the

hGnRH-II promoter Overexpression of p65 in both

TE-671 and JEG-3 cells led to a dramatic decrease in

hGnRH-II promoter activities and endogenous gene

expression Moreover, differential regulation of the

GnRH-II gene in two different GnRH-II-expressing

human cell lines was observed in assays involving

overexpression of RARa and cotransfection of RARa

and RXRa In the studies, an increase in promoter

activity was found in placental JEG-3 cells, but no effect

could be observed in neuronal TE-671 cells

(Bu)2cAMP

Chen et al revealed the presence of an 8 bp

palin-dromic cAMP response element consensus site

(TGACGTCA, )67 to )60 bp) in the hGnRH-II

pro-moter [55] In TE-671 cells, treatment with 1 mm

(Bu)2cAMP for 12–48 h strongly upregulated GnRH-II

gene expression, and this was verified by RT-PCR and

immunofluorescence staining An increased

concentra-tion of GnRH-II peptides was also observed in the cell

medium after the treatment Moreover, strong

induc-tion of promoter activities of the hGnRH-II gene in

response to 1 mm (Bu)2cAMP was found in

transfec-tion studies on the hGnRH-II promoter construct

cou-pled to luciferase

Self-regulation of GnRH gene

The role of GnRH in the regulation of synthesis and

secretion of gonadotropins in the pituitary is well

known Recent studies have revealed GnRH as an

autocrine and paracrine regulator of gonadotropins in

the hypothalamus and ovary It has been demonstrated

that GnRH-I gene expression is regulated by itself

through an ultrashort loop feedback mechanism in rat

hypothalamus [56] and ovarian cells [57] Also, it has

been shown that the GnRH-I and GnRH-II genes are

differentially regulated by themselves

In vitro studies using GT1-7 cells and hypothalamic

tissue cultures, and in vivo studies in a rat ovary

model, showed that GnRH-I treatment inhibited the

expression and secretion of GnRH-I [56,58–61] In

hGLCs, GnRH-I has been shown to be regulated by its

own ligand [62] Treatment with a GnRH-I analog

(leuprolide) produced a biphasic effect on GnRH-I

mRNA levels, depending on the concentration of

treat-ment Low concentrations of leuprolide (10)11 and

10)10m) resulted in upregulation of GnRH-I gene

expression, whereas high concentrations (10)8 and

10)7m) led to gene repression This type of biphasicregulation has also been shown in immortalized hypo-thalamic GT1-7 cells [63] and human OSE cells [64].Treatment with an antagonist (antide) prevented thisbiphasic effect in OSE cells, proving the specificity ofthe response Moreover, intracerebroventricular injec-tion of a GnRH-I analog into the lateral ventricle ofrat brain resulted in a considerable decrease in GnRH-ImRNA levels, in a dose- and time-related manner, asdetected in the POA [56] For GnRH-II, treatmentwith different concentrations of the homologous ligandand GnRH-II analog (10)11 and 10)7m) in hGLCsresulted in a large decrease in GnRH-II mRNA levels

In a human endometrial cell line, Ishikawa, ment with GnRH-I increased GnRH-I expression in atime-dependent manner, but did not cause any change

treat-in GnRH-II mRNA levels [65] These data showed thatthe GnRH-I and GnRH-II genes are differentially regu-lated by their own ligands, suggesting the differentialregulation of the two forms of GnRH in differentstages of the estrous cycle The exact mechanism forthis differential regulation is unclear Kang et al sug-gested the possibility of different characteristics ofbinding of the two forms of GnRH to their receptor,which might lead to different conformations of thereceptor [62] The ligand-specific conformation mighttherefore lead to differential coupling to G-proteinsand⁄ or different intracellular cellular pathways, even-tually leading to differential regulation of GnRH-I andGnRH-IIgene expression

GnRH-I(1–5) is a pentapeptide that comprises thefirst five amino acids of GnRH-I It is a processed pep-tide formed by cleavage of the Try5-Gly6 bond by azinc metalloendopeptidase, EC 3.4.24.15 (EP24.15)[65] Wu et al demonstrated that GnRH-I(1–5) stimu-lated GnRH-I mRNA expression in neuronal GT1-7cells through a different pathway from that used bythe parent peptide GnRH-I [66] In Ishikawa cells,GnRH-I(1–5) was found to have no effect on GnRH-ImRNA expression, but induced GnRH-II mRNAexpression Baldwin et al suggested that the differ-ences between the actions of GnRH-I and its metabo-lite GnRH-I(1–5) on the regulation of the GnRH-I andGnRH-II genes could be caused by the two peptidesacting through different GnRH receptors [65]

Gonadotropins

Gonadotropins, including FSH and LH, are secreted

by gonadotropes of the pituitary under the control ofGnRH A third gonadotropin that is also present inhumans is hCG, which is produced in the placenta

during pregnancy Like GnRHs, gonadotropins have

been shown to differentially regulate the two GnRH

genes via stimulation of cAMP production and

activa-tion of PKA [1] Through these pathways,

gonadotro-pins may regulate the ratio between GnRH-I and

GnRH-II, leading to distinct spatial expressions of the

two hormones

In GT1-7 cells, gonadotropins are downregulators of

the GnRH-I gene Lei and Rao showed that GnRH-I is

coexpressed with LH⁄ hCG receptor in rat POA and

GT1-7 cells [67] Treatment of GT1-7 neurons with

LH or hCG resulted in a decrease in steady-state

GnRH-I mRNA levels This decrease was found to be

dose- and time-dependent, and required the presence

of cellular LH⁄ hCG receptors The same group then

investigated the signaling pathway and factors involved

in the action of hCG [40] A cAMP analog,

8-bromo-cAMP, was reported to mimic the downregulation

action of hCG, and application of a PKA inhibitor

H89, but not a PKC inhibitor, blocked the action of

hCG and that of the cAMP analog These findings

suggested that PKA signaling and transcription factors

such as CREB, Fos and Jun are probably involved in

transcriptional inhibition of GnRH gene expression by

hCG in GT1-7 cells Later, the group further extended

the study to investigate the cis-acting elements and

trans-acting proteins involved in the inhibition by hCG

[41] Deletion analysis revealed that the region between

)126 and )73 bp is important for the hCG inhibition

Within this region, the)99 to )94 bp region contained

an imperfect AP-1 site, and the )91 to )81 bp region

contained two AT-rich sites ()91 to )87 bp and )85

to )81 bp) Also, southwestern blots showed that a

110-kDa protein and a 95-kDa protein bound to the

)126 to )73 bp region Mutagenesis of the AT-rich

site, but not the AP-1 site, resulted in complete loss of

the inhibitory effect of hCG and also DNA binding of

the 95 kDa protein However, supershift assays have

not yet been able to determine the identity of the

95 kDa protein

The actions of gonadotropins on GnRH-I and

GnRH-II gene expression have been shown to be

diverse In the past, treatment with hCG (1 IUÆmL)1)

in hGLCs did not affect the expression level of the

rGnRH-I gene, but decreased GnRH receptor mRNA

levels [68] However, later studies showed that

treat-ment with FSH or hCG in hGLCs resulted in a

decrease in GnRH-I mRNA levels, but a significant

dose-dependent increase in GnRH-II mRNA levels

[62] Recently, it was found that GnRH-II mRNA

lev-els were significantly reduced following FSH or LH

treatment (100 ngÆmL)1and 1000 ngÆmL)1) for 24 h in

the two IOSE cell lines (IOSE-80 and IOSE-80PC) and

three ovarian cancer cell lines (A2780, BG-1 and AR-3) [69] In contrast, treatment with either FSH or

OVC-LH had no effect on GnRH-I mRNA levels in the celllines employed These findings suggested thatgonadotropins regulate the two forms of GnRH differ-ently in the ovary

Steroid hormones

The pulsatile secretion of GnRH from the lamic neurons regulates the synthesis and release ofgonadotropins in the pituitary The gonadotropinsthen regulate both steroidogenesis and gametogenesis.The gonadal steroid hormones, which are key regula-tors of reproduction, in turn act tightly to regulateGnRH-I and GnRH-II synthesis and release through anegative feedback system between the gonads and thebrain The effects of 17b-estradiol (E2) and progester-one (P4) and their receptors on GnRH gene expressionhave been well studied

hypotha-Previously, a number of studies found an absence ofsteroid receptors in GnRH neurons [70] It wasbelieved that GnRH neurons synapse with other neu-rons that act as potential gonadal steroid-sensitiveinterneurons to modulate GnRH neurons through

a number of neurotransmitters and neuropeptides [71].However, recent studies revealed the expression ofdifferent steroid receptors in various hypothalamic andovarian cell lines [72,73] Belsham et al suggested thatthe apparent absence of steroid receptors was probablydue to the scarcity and scattered distribution of GnRHneurons, or to the fact that only specific subgroups ofGnRH neurons may contain steroid receptors There-fore, steroid receptors were not detected [71] Also, itmight be due to limitations in the sensitivity of thedetection methods These steroid receptors, upon form-ing complexes with their specific steroid hormoneligands, act as intracellular transcription factors, andexert their effects on the expression of GnRH genes

EstrogenThe discovery of the estrogen response element ()441

to)428 bp) in the hGnRH-I gene [74] and the presence

of both forms of nuclear estrogen receptors (ERa andERb) in hypothalamic and ovarian cell lines, GT1-7and hGLCs [11,75,76], suggested the possible involve-ment of estrogen (E2) in regulation of the GnRH-Igene

Inconsistent results have been reported regardingthe regulation of GnRH-I gene expression by E2 inthe hypothalamus This is because GnRH-I is regu-lated differently at different stages of the estrous