However, in contrast to GnRH-I, GnRH-II is expressed at significantly higher levels outside the Keywords cross-talk; extrapituitary; GnRH; GnRH receptor; MAPK; metastasis; pituitary; rece
Trang 1Gonadotropin-releasing hormone: GnRH receptor signaling
in extrapituitary tissues
Lydia W T Cheung and Alice S T Wong
School of Biological Sciences, University of Hong Kong, China
Introduction
The hypothalamic gonadotropin-releasing hormone
(GnRH) is a decapeptide that plays a critical role in
the regulation of reproduction GnRH-I
(pGlu-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-Gly-NH2) is the first
GnRH isoform discovered in mammalian brain Its
major role is to stimulate pituitary secretion of
gonadotropins, luteinizing hormone and follicle-stimu-lating hormone, which in turn stimulate the gonads for steroid production Subsequently, a second iso-form of GnRH (His5, Trp7, Tyr8) (GnRH-II) has been isolated from chicken brain It is also highly conserved among vertebrates, including mammals [1] However, in contrast to GnRH-I, GnRH-II is expressed at significantly higher levels outside the
Keywords
cross-talk; extrapituitary; GnRH; GnRH
receptor; MAPK; metastasis; pituitary;
receptor tyrosine kinase; signaling; tumor
Correspondence
A S T Wong, School of Biological
Sciences, University of Hong Kong, 4S-14
Kadoorie Biological Sciences Building,
Pokfulam Road, Hong Kong, China
Fax: +852 2559 9114
Tel: +852 2299 0865
E-mail: awong1@hku.hk
(Received 14 April 2008, revised 28 May
2008, accepted 11 June 2008)
doi:10.1111/j.1742-4658.2008.06677.x
Gonadotropin-releasing hormone (GnRH) has historically been known as
a pituitary hormone; however, in the past few years, interest has been raised in locally produced, extrapituitary GnRH GnRH receptor (GnRHR) was found to be expressed in normal human reproductive tissues (e.g breast, endometrium, ovary, and prostate) and tumors derived from these tissues Numerous studies have provided evidence for a role of GnRH
in cell proliferation More recently, we and others have reported a novel role for GnRH in other aspects of tumor progression, such as metastasis and angiogenesis The multiple actions of GnRH could be linked to the divergence of signaling pathways that are activated by GnRHR Recent observations also demonstrate cross-talk between GnRHR and growth fac-tor recepfac-tors Intriguingly, the classical Gaq–11-phospholipase C signal transduction pathway, known to function in pituitary gonadotropes, is not involved in GnRH actions at nonpituitary targets Herein, we review the key findings on the role of GnRH in the control of tumor growth, progres-sion, and dissemination The emerging role of GnRHR in actin cytoskele-ton remodeling (small Rho GTPases), expression and⁄ or activity of adhesion molecules (integrins), proteolytic enzymes (matrix metalloprotein-ases) and angiogenic factors is explored The signal transduction mecha-nisms of GnRHR in mediating these activities is described Finally, we discuss how a common GnRHR may mediate different, even opposite, responses to GnRH in the same tissue⁄ cell type and whether an additional receptor(s) for GnRH exists
Abbreviations
EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; ERK, extracellular signal-related kinase; FAK, focal adhesion kinase; FGF, fibroblast growth factor; GnRH, gonadotropin-releasing hormone; GnRHR, gonadotropin-releasing hormone receptor;
JNK, Jun N-terminal kinase; MAPK, mitogen-activated protein kinase; MMP, matrix metalloproteinase; NF-jB, nuclear factor kappa B; PI3K, phosphatidylinositol 3-kinase; PKC, protein kinase C; Pyk2, proline-rich tyrosine kinase 2; RTK, receptor tyrosine kinase; uPA, urokinase-type plasminogen activator; VEGF, vascular endothelial growth factor.
Trang 2brain and is particularly abundant in the kidney,
bone marrow, and prostate [2] This leads to the
speculation that GnRH-II may have distinct
physio-logical functions from those of GnRH-I In line with
this is the observation that although GnRH-II can
stimulate gonadotropin secretion, its efficiency is
much lower than that of GnRH-I (only about 2% of
that of GnRH-I) [3] This suggests that the primary
role of GnRH-II is not in the regulation of
gonado-tropin secretion Instead, this peptide has been shown
to act as a neuromodulator [4] The exact actions of
GnRH-II in peripheral tissues are not entirely
under-stood, but this is certainly an important topic for
investigation which may offer an opportunity to
eluci-date the undisclosed complexity of GnRH
In this minireview, we will focus on recent progress
in understanding the roles of GnRH-I and GnRH-II
in extrapituitary tissues, in particular its emerging
role in tumor growth, invasion, and metastasis We
will also describe the molecular mechanisms underlying
these effects, focusing on the roles of proteolysis,
adhesion, and signaling, as well as our still-emerging
understanding of receptor cross-talk with other
pathways Finally, we will discuss two important
outstanding questions in the field regarding what might
distinguish the different responses to the same ligand
(GnRH) and whether an additional receptor(s) for
GnRH exists in humans
Localization of GnRH receptor (GnRHR)
in peripheral reproductive tissues
The initial interest in extrapituitary GnRHR stemmed
primarily from observations in the 1980s that GnRH
analogs can inhibit the growth of nonpituitary tumor
cell lines [5] Soon after this, a functional type I
GnRHR was demonstrated in a variety of normal
human reproductive tissues (e.g breast, endometrium,
ovary, and prostate) and tumors derived from these
tissues
In the ovary, GnRHR mRNAs are expressed in
granulosa-luteal cells, and increased expression of
GnRHR correlates with follicular growth and
develop-ment [6] GnRHR binding has been demonstrated in
luteinized granulosa cells, late follicles and developing
corpora lutea, but not in primordial, early antral and
preovulatory follicles [7,8] This stage-specific
expres-sion of GnRHR in the human granulosa and luteal
cells suggests a role for GnRH in the regulation of
ovarian physiology, particularly ovulation, follicular
atresia and luteolysis The presence of GnRHR protein
and mRNA has also been demonstrated in human
ovarian tumor specimens, ovarian cancer cell lines and
their tissue of origin, ovarian surface epithelium [9,10] Interestingly, levels of GnRHR seem to be associated with cancer grading and have been reported to be elevated in advanced stage (stages III and IV) as compared to early stage (stages I and II) ovarian carcinomas [11] Our recent findings that GnRH can promote the motility and invasiveness of ovarian can-cer cells further corroborate the view that GnRH may play a crucial role in tumor progression⁄ metastasis [12,13], and these findings will be discussed in a later section
Using [125I][d-Trp6]GnRH, specific receptor binding has been detected in membranes from 24 of 31 (77%) endometrial carcinomas and from three of 13 (23.1%) nonmalignant human endometrial specimens [14] GnRHR mRNA has been clearly detected in surgical endometrial carcinoma specimens and endometrial carcinoma cell lines [15,16] As with normal myome-trium, most benign neoplasms studied thus far, including uterine leiomyoma, also possess GnRHR [17]
Early studies showed that the human placenta con-tains specific binding sites for GnRH that interact with GnRH agonists and antagonists [18] Later on, GnRHR was localized to the cytotrophoblast and syncytiotrophoblast cell layers [19,20] Temporal expression of GnRHR in the placental cells at different weeks of gestation has been observed, in parallel with the time-course of chorionic gonadotropin secretion during pregnancy [21], suggesting that the expression
of the receptor is a function of pregnancy stage The presence of GnRHR has been demonstrated in numerous human breast cancer cell lines and tumor biopsy specimens [22–24] GnRHR was immunolocal-ized in the cytoplasm in 37 of 58 (64%) invasive ductal carcinoma cases [23] The expression of GnRHR in normal human breast tissue is still controversial, but the sample size may have been too small to allow any definite conclusion [22,25]
GnRHR is also present in prostate cancer cells, as shown by radioligand-binding studies, PCR, and western blotting analysis [26,27] GnRHR immunore-activity is localized to the luminal and basal epithelial cells in benign and malignant prostate tissues In this study, the relative GnRHR mRNA levels showed a wide range of individual differences that were unrelated to the histological grades of the 16 cases [27] There does, however, appear to be significantly higher expression of GnRHR in hormone-refractory prostate carcinoma than in other types of prostate tumor (n = 80) [28] Although these extrapituitary GnRHRs share the same cDNA nucleotide sequence and encode tran-scripts and proteins of the same size as the pituitary
Trang 3GnRHR [20,26,29], they also differ in several ways.
First, cell surface receptor expression in extrapituitary
sites is low as compared to that of the pituitary
[15,27] This may underlie the greater effect of the
GnRHR ligands on the gonadotropes Second, there
are at least two classes of GnRHR: one has high
affin-ity [with nanomolar dissociation constants (Kd)] for
GnRH, and one has low affinity (with micromolar Kd
values) for GnRH The high-affinity GnRH-binding
sites are commonly regarded as being the same as the
GnRHR of the pituitary gland Whereas in most of
the reported cases, both the low-affinity and
high-affin-ity GnRHR have been found in extrapituitary tissues
[30–33], in some cases, only low-affinity GnRHR could
be detected [10,18,34], and in others, e.g in
endome-trial cancers and nonmalignant endomeendome-trial specimens,
only the high-affinity GnRHR has been demonstrated
[14] The exact role of each of these receptors and the
implications of differential levels of expression remain
to be elucidated
Functions of GnRH-I and GnRH-II in
cancers
Tumor growth
Over the last two decades, both GnRH agonists and
antagonists have been widely used as therapeutics in
treating sex steroid-dependent tumors The majority
of these GnRH analogs, when given continuously,
inhibit gonadotropin synthesis and secretion via
downregulation of the pituitary GnRHRs This
indi-rect mechanism of action has provided the rationale
for the use of GnRH analogs in the treatment of
hor-mone-dependent tumors for many years Only since
the detection of GnRHR in extrapituitary tissues has
there been increasing interest in its direct action on
tumor cells
GnRH-I analogs have direct antiproliferative effects
on ovarian cancer cells, which is linked to the
disrup-tion of the cell cycle at G0⁄ G1 [31,35,36] On the other
hand, several independent in vitro studies failed to
demonstrate significant growth inhibition by GnRH-I
agonists, even at fairly high concentrations (micromolar
range) [37,38] In fact, a biphasic impact of GnRH-I
agonists on growth has been reported: whereas GnRH-I
agonists at high dose (1 lm) inhibit cell proliferation
in vitro, cells treated with agonists at low dose (10 nm)
show significant growth stimulation [39] Further
studies demonstrated that nanomolar concentrations of
GnRH-I agonists also increase cell survival under
multiple stress conditions, including DNA
replication-specific cytotoxic agents and UV radiation [40]
GnRH-II has antiproliferative effects on ovarian cancer cells [41–43] Although it has been suggested that this effect of GnRH-II is mediated through the type I GnRHR [43], there are other findings implicating a type I GnRHR-independent action [41,42]
It is interesting to note that although both GnRH-I agonists and antagonists exert antiproliferative effects, the effects of GnRH-I antagonists are stronger than those of the agonists [44] This difference has also been seen in an in vivo model, which demonstrates a signifi-cant inhibition of tumor growth by GnRH-I antago-nists but not GnRH-I agoantago-nists [45] The advantage of GnRH antagonists over the agonistic peptides is prob-ably due to the fact that they inhibit the secretion of gonadotropins and reduce sex steroid levels immedi-ately after application, thus achieving rapid therapeutic effects, whereas repeated exposure to agonistic agents
is required to induce functional desensitization of the gonadotropes [46]
Treatment of human endometrial cancer cells (cell line Ishikawa) with the GnRH-I antagonist SB-75 results in growth inhibition, mainly due to the Fas⁄ Fas ligand-mediated apoptotic pathway, whereas GnRH-I agonists have no effect on the same cell line [15,47,48] Another endometrial carcinoma cell line, HEC-1A, also exhibits differential responses to different GnRH agon-ists and antagonagon-ists [15,30,36,48] GnRH-II has been shown to have antiproliferative effects on endometrial carcinoma cells [41] The effects of GnRH-I are abrogated after type I GnRHR knockout [36], whereas those of the GnRH-I antagonist cetrorelix and of GnRH-II persist [41] These findings suggest that the antiproliferative effects of cetrorelix and GnRH-II are not mediated through the type I GnRHR
GnRH-I has been demonstrated to have antiprolifer-ative effects on prostate cancer cells [49–51], except in one in vivo study [52] This antiproliferative effect appears to be independent of the androgen receptor status of the prostate carcinoma cells, as both andro-gen-sensitive LNCaP cells and androgen-resistant DU-145 cells remain sensitive to GnRH [49,50] Acti-vation of GnRHR may mediate these effects via direct induction of apoptosis through caspase activation [53] Compatible with a role for GnRH in survival at low doses, an enhancing effect of GnRH was observed when cells were treated with a low concentration (100 pm) of GnRH-I agonist [54] GnRH-II was shown
to have an antiproliferative effect on DU-145 cells and growth-stimulatory effect on TSU-Pr1 cells, but the type I GnRHR was not involved [55]
The influence of GnRH on the growth of human breast cancer cells was first studied with MCF-7 cells [56], and both in vitro and in vivo proliferation of
Trang 4breast cancer cells could be inhibited by both agonistic
and antagonistic analogs of GnRH [57,58] However,
higher efficiency of GnRH antagonists in growth
inhi-bition than that of GnRH agonists has been reported
[24,58]
Invasion and metastasis
The observation that GnRH controls tumor growth
suggests a regulatory role for this peptide in the
meta-static behavior of cancer cells This hypothesis is
sup-ported by studies showing that GnRH-I and GnRH-II
can affect the expression of several extracellular
matrix-degrading enzymes in human extravillous
cyto-trophoblasts and decidual stromal cells to facilitate
implantation [59,60] However, its potential role in
cancer metastasis has just begun to be revealed
Metastasis is a complex phenomenon that requires
several specific steps, such as decreased adhesion,
increased motility, and proteolysis The effects of GnRH
in tumor metastasis are mediated through the regulation
of adhesion molecules, Rho GTPases, and two families
of metastasis-related proteinases, the matrix
metallopro-teinases (MMPs) and the urokinase-type plasminogen
activator (uPA) system, at several levels: mRNA
transcription, secretion, and proenzyme activation
The ability of GnRH to regulate metastasis was first
reported in melanoma cells [61] High doses of GnRH-I
analog, at micromolar concentrations, significantly
reduces the ability of melanoma cells to invade
and migrate [61] Preliminary data (R M Moretti,
M Monagnani Marelli, J C van Groeninghen, M
Motta & P Limonta, unpublished results, 2003) indicate
that this inhibitory action is due to the effects of
integrins and MMPs [62]
We were the first to report possible metastatic
activ-ity of GnRH-I in tumors of the female reproductive
tract [12] GnRH-I exerts a biphasic effect on cellular
migration and invasion: whereas lower (nanomolar)
concentrations of the GnRH-I agonist stimulate
cellu-lar migration and invasion in a dose-dependent
man-ner, high (micromolar) concentrations are not as
efficient This proinvasive effect is mediated through
activation of metastasis-related proteinases, in
particu-lar MMP-2 and MMP-9 [12] Moreover, GnRH-I is
able to transactivate the MMP-2 and MMP-9
promot-ers, which means that GnRH can be considered to be
a new member of MMP-2 and MMP-9 transcriptional
modulators Like GnRH-I, native GnRH-II and its
synthetic analog also induce a similar biphasic
regula-tion of ovarian cancer invasion [13] The finding that
small interfering RNA-mediated downregulation of
type I GnRHR completely reversed the effects of both
GnRH-I and GnRH-II on cell invasion supports the view that the same receptor, type I GnRHR, is essen-tial for the effects of GnRH-I and GnRH-II in ovarian cancer cells
The decrease in uPA activity of cytosol from Dun-ning R3327H rat prostate tumors after treatment with GnRH-I analogs suggests that GnRH may be an important factor in reducing the invasiveness of pros-tate cancer [63] High doses of GnRH-I agonists and antagonists have been reported to attenuate the invading capacity of both androgen-dependent and androgen-independent prostate cancer cells by modu-lating E-cadherin-mediated cell–cell contacts and pro-duction of uPA and its inhibitor (plasminogen activator inhibitor-1) [64–66] GnRH has also been shown to regulate cell motility through its interaction with the small GTPases Rac1, Cdc42, and RhoA, which are involved in the regulation of actin polymer-ization [67]
Up to now, there has been only one study, by Von Alten et al., investigating the role of GnRH in breast cancer metastasis, using a coculture system with human osteosarcoma cells to analyze tumor cell invasion to bone [68] The consequences of GnRHR activation are complex and appear to be cell context dependent: whereas treatment of cells with the GnRH-I agonist triptorelin, the GnRH-II agonist [d-Lys6]GnRH-II and the GnRH-I antagonist cetrorelix decreases the invasion rate in most breast cancer cell lines, these agents have no significant effect in the GnRHR-positive MDA-MB-435 cells [68] Further investigations are required to elucidate the reason why the MDA-MB-435 cell line reacts differently
Organ-specific homing and colonization of cancer cells are important and interesting features of metasta-sis A role for GnRH has also been suggested in the regulation of the immune response and metastasis GnRH-I and GnRH-II are expressed in human normal and cancerous T-cells GnRH triggers laminin receptor gene expression, adhesion to laminin, in vitro chemo-taxis, and in vivo homing to specific organs [69]
Angiogenesis Angiogenesis is crucial to a number of physiological and pathological processes, such as reproduction, development, and tissue repair, as well as tumor growth and metastasis Vascular endothelial growth factor (VEGF) is implicated as the most important angiogenesis inducer, because of its potency in a variety of normal and tumor cells Other angiogenic factors include fibroblast growth factor (FGF), plate-let-derived growth factor and the angiopoietin family
Trang 5The effect of GnRH on angiogenesis in the ovary, in
which this neovascularization is necessary for follicular
and luteal function, has been demonstrated A recent
in vivostudy using rats revealed that an application of
the GnRH-I agonist leuprolide acetate decreases the
protein expression of VEGF and angiopoietin-1 and
their receptors in ovarian follicles, and that this can be
reversed by coinjection of the GnRH antagonist antide
[70] A similar inhibitory effect on angiogenesis can be
observed in marmosets injected with the GnRH-I
antagonist antarelix [71] However, VEGF mRNA
expression is unaffected by the treatment The clinical
response of uterine shrinkage after GnRH analog
treatment and a pathological role of FGF-2, VEGF
and platelet-derived growth factor in uterine
leiomy-oma growth and vascularization has also been
sug-gested [72] Considering that angiogenesis is an
important process in many human cancers, it would be
very interesting to determine whether GnRH also plays
a key role in tumor angiogenesis
Intracellular signal transduction
Upon GnRH binding, GnRHR undergoes a conforma-tional change and stimulates a unique G-protein Inter-estingly, the classical Gaq–11-phospholipase C signal transduction pathway, which is known to operate in the pituitary, is not involved in the antitumor activity
of GnRH analogs Rather, GnRHRs couple to Gaiin these tumors and result in the activation of several downstream signaling cascades [73,74], such as mito-gen-activated protein kinase (MAPK), phosphatidyl-inositol-3-kinase (PI3K), and nuclear factor kappa B (NF-jB) signaling The GnRH-induced signaling path-ways in extrapituitary tissues are shown schematically
in Fig 1
Fig 1 Schematic representation of GnRHR signaling in extrapituitary tissues Binding of GnRH to GnRHR triggers several intracellular signal-ing cascades and cross-talk with mitogenic signalsignal-ing, dependsignal-ing on the cell context Some of these signalsignal-ing modules can transduce extra-cellular signals to the nucleus and thereby regulate genes that are involved in cell growth, metastasis, or survival Arr, b-arrestin; CREB, cAMP response element-binding protein; FGFR, fibroblast growth factor receptor; HB-EGF, heparin-binding EGF; IjB, inhibitory factor
kap-pa B; IGFR, IGF receptor; MEK, mitogen-activated protein kinase kinase; MLK3, mixed-lineage kinase 3; PTP, protein tyrosine phosphatase; Sos, son of sevenless; TNF-a, tumor necrosis factor alpha.
Trang 6The major MAPK cascades include extracellular
sig-nal-regulated kinase (ERK), Jun N-terminal kinase
(JNK), and p38 MAPK Many studies have shown
that the MAPK pathway is critical for GnRH
activi-ties, which provides an important link for the
trans-mission of signals from the cell surface to the nucleus
Activation of MAPK by GnRH involves distinct
upstream pathways in generating tissue-specific and
cell-specific signaling (Fig 1) This can occur at
differ-ent levels via differdiffer-ent mechanisms: (a) second
messen-gers [protein kinase C (PKC) and cAMP] [26];
(b) receptor-interacting proteins such as Src and
b-zarrestin [53,75]; and (c) upstream kinases such as
MAPK⁄ ERK kinase and PI3K [53,76] For example,
GnRH-I induces apoptosis in DU-145 prostate cancer
cells via JNK, which is activated through two
indepen-dent mechanisms [53] Activation of the pathway is
dependent on c-Src with concomitant decrease in Akt
activity, and the combination of these two events
relieves the inhibition of the upstream activator of
JNK, MLK3 [53] (Fig 1) Interestingly, although
ERK1⁄ 2 is phosphorylated through epidermal growth
factor receptor (EGFR) under the same conditions,
this pathway is not involved in the apoptotic effects
These findings demonstrate that activation of the two
MAPKs, which lead to distinct physiological
out-comes, is separated already at the upstream levels In
the ovarian cancer cell line CaOV-3, prolonged
stimu-lation of ERK1⁄ 2 through Shc and son of sevenless is
required for GnRH-I-mediated growth inhibition [76]
Consistent with the fact that sustained activation of
ERK1⁄ 2 is often correlated with cell cycle progression,
GnRH-I-induced growth inhibition is attributed to G1
arrest [76] Moreover, the signaling cascade was shown
to be initiated by Gbc, supporting the notion that the
post-GnRHR signaling cascade in extrapituitary cells
is different from that in pituitary cells GnRH-induced
MAPK activation has also been shown in another
ovarian cancer cell line, OVCAR-3 Both ERK1⁄ 2 and
p38 MAPK mediate the antiproliferative effects of
GnRH-I and GnRH-II in a PKC-dependent manner
[43,77] GnRH-II induces the activation of activator
protein-1 transcription factor via p38 MAPK,
suggest-ing a potential role of activator protein-1 in ovarian
cancer cell growth [77] The JNK pathway also drives
tumor invasion and migration in ovarian cancer cells
[12], but the activation mechanism(s) remains to be
elucidated
Temporal and spatial differences in cellular signaling
may have significant phenotypic manifestations [78,79]
For example, sustained activation of ERK1⁄ 2 has been
implicated in nerve growth factor-mediated neuronal differentiation of PC12 cells, whereas a rapid and transient activation is associated with growth factor-mediated proliferation of PC12 cells [80] Thus, the duration of kinase activation seems to be a major determinant of signal outcome We have shown differ-ential regulation of ERK1⁄ 2, p38 MAPK, and JNK
by GnRH-I with sustained signaling through the JNK pathway in ovarian cancer cells [12] Consistently, GnRH-stimulated MMP-2 and MMP-9 expression, secretion and cell invasion were attenuated by specific inhibition of JNK but not of ERK1⁄ 2 and p38 MAPK, suggesting that prolonged activation of JNK may contribute to a more invasive phenotype Strong and sustained activation of MAPK has been reported
to be necessary for its cytoplasm-to-nucleus transloca-tion, and thereby contributes to the regulation of gene expression [79,81] It will be interesting to see whether sustained activation of JNK leads to its nuclear trans-location, which is required for GnRH-stimulated cell invasion The JNK pathway targets multiple transcrip-tion factors, including c-Jun, c-Fos, ATF and PEA, and putative binding sites for these DNA-binding pro-teins are present in the MMP promoters [82] Whether these putative regulatory elements participate in the GnRH-dependent activation of the MMP-2 and MMP-9 genes remains to be determined
Cross-talk with mitogenic signaling Cross-talk between cell surface receptors, which has been recognized as a mechanism capable of generating signal diversity, is now receiving further interest Figure 1 illustrates the cross-talk between GnRHR and receptor tyrosine kinases (RTKs) For instance, GnRH causes transactivation of RTKs, such as EGFR [75,78,83] MMP-2 and MMP-9 seem to be essential for GnRH-induced EGFR activation by cleavage of the heparin-binding epidermal growth factor (EGF) precursor [84] Transactivation of EGFR has been shown to activate ERK1⁄ 2, as GnRH-induced ERK1⁄ 2 phosphorylation can be abolished in the pres-ence of the EGFR inhibitor AG1478 [53,78] However, the biological importance of ERK1⁄ 2 activation in response to this cross-talk still remains elusive
Negative cross-talk between GnRHR and growth factor receptors has also been described For instance, the antiproliferative effects of GnRH-I and GnRH-II agonists are mediated through attenuation of EGFR signaling in many reproductive tumor cells [57,66,85– 87] In prostate cancer cells, cetrorelix is able to coun-ter EGFR-dependent adhesive signaling through
a PKC-dependent mechanism [66] Activation of
Trang 7GnRHR appears to mediate these effects via activation
of phosphotyrosine phosphatase, thereby reducing
EGF-induced EGFR autophosphorylation, resulting in
downregulation of mitogenic signal transduction and
cell proliferation [85,86,88] A negative regulatory
interaction between GnRHR and mitogenic signaling
pathways has also been reported in human prostate
cancer cells via insulin-like growth factor GnRH-I
agonists inhibit expression of the insulin-like growth
factor receptor, receptor tyrosine phosphorylation, and
the subsequent downstream activation of Akt [89–91]
Another example is FGF-2 GnRH analog treatment
has been shown to block cell proliferation and
inva-sion induced by FGF-2 stimulation [65]
PI3K
The PI3K signaling pathway and its downstream
target Akt (also named protein kinase B) has been
implicated in promoting cell survival, proliferation,
and invasion In uterine leiomyomas, the GnRH-I
ago-nist leuprolide acetate causes a significant reduction in
PI3K⁄ Akt activity and inhibits the expression of the
antiapoptotic proteins (c-FLIP and PED⁄ PEA15),
thereby inducing apoptosis [92] In the SKOV-3
ovar-ian cancer cell line, GnRH-I and GnRH-II interfere
with activation of the PI3K⁄ Akt cascade, and this is is
associated with the inhibitory effects of GnRH on cell
invasion [13]
Although PI3K⁄ Akt and MAPK are two parallel
pathways in some cell types, they are two related
path-ways in the mediation of GnRH actions, as inhibition
of PI3K⁄ Akt can alter the activation of MAPK For
instance, in prostate cancer cells, stimulation of
PI3K⁄ Akt releases mixed-lineage kinase 3, which in
turn activates the JNK pathway, and this positive
reg-ulation is important for the proapoptotic effect of
GnRH-I (Fig 1) [53] PI3K⁄ Akt is also an upstream
kinase of ERK1⁄ 2, and EGFR transactivation by
GnRH-I may be required for the activation of this
cascade [75,93]
Other signaling pathways
Activation of NF-jB is important for the protection
against apoptosis in ovarian tumors induced by the
GnRH-I agonist tiptorelin [94] The effect is probably
mediated by the Gai-coupled GnRHR, and receptor
activation causes nuclear translocation of NF-jB [94]
Unlike the other signaling pathways studied, the
GnRH-I-induced NF-jB activation appears to be
inde-pendent of the cross-talk between GnRHR and growth
factor signaling, as treatment with phosphatase
inhibi-tor has no effect on the activation of NF-jB [94] It has also been shown that GnRH-I suppresses interleu-kin-8 expression via attenuation of tumor necrosis fac-tor alpha-induced NF-jB signaling in endometriotic stromal cells (Fig 1) [95] These data suggest that modulation of cytokine signal transduction by GnRH may be one of the mechanisms contributing to its growth-inhibitory effect
The non-RTKs focal adhesion kinase (FAK) and proline-rich tyrosine kinase 2 (Pyk2) are the predomi-nant mediators of integrin signaling GnRHR has been shown to signal through these molecules, suggesting a role for GnRH in cytoskeletal reorganization In human endometrial cancer cells (HEC-1A), b3 -integrin-dependent activation of FAK is associated with the inhibitory effects of GnRH-I and GnRH-II on growth [96] Leiomyoma regression induced by GnRH-I agon-ists has been suggested to be mediated, at least in part, through a mechanism involving suppression of FAK [97] Maudsley et al demonstrated a novel signaling cascade of GnRHR that functionally antagonizes the actions of testosterone and inhibits prostate tumor growth [98] GnRH controls the tyrosine phosphoryla-tion status of the focal adhesion proteins Pyk2 and Hic-5 This alteration of the focal adhesion dynamics then results in nuclear translocation of the androgen receptor, which renders it transcriptionally inactive [98]
Mechanisms underlying the diverse responses to GnRH action
As discussed earlier, GnRH and its agonists have a dual and biphasic action: whereas low concentrations (0.1–10 nm) of GnRH stimulate cell proliferation, migration and invasion in a dose-dependent manner, high concentrations (100 nm to 1 lm) inhibit these functions [12,13,39] Moreover, the same dose of GnRH can elicit completely opposite responses in cells derived from the same tissue We demonstrated that in two human ovarian cancer cell lines, OVCAR-3 and SKOV-3, GnRH-I and GnRH-II induce invasion of OVCAR-3 cells, but inhibit the invasiveness of
SKOV-3 cells [1SKOV-3] A similar difference has been found in the effects of GnRH on cell proliferation and cell migra-tion in the prostate carcinoma cell lines TSU-Pr1 and DU-145 [67] Whereas GnRH-I and GnRH-II stimu-lated cell proliferation, induced actin cytoskeleton remodeling and promoted migration in TSU-Pr1 cells, they were inhibitory in DU-145 cells [67] The observa-tion that GnRH-I and GnRH-II have no significant effect in cell lines with type I GnRHR depletion indi-cates that the type I GnRHR is indispensable for the
Trang 8effects of both GnRH-I and GnRH-II [13,67] Thus,
one intriguing question is how a common GnRHR
may mediate different, even opposite, responses to
GnRH in the same cell type⁄ tissue The reasons are
unknown, but several possibilities (as summarized in
Table 1) can be envisaged First, the treatment
condi-tion may be one determinant of the outcome The
pulsatility of GnRH release is necessary for the
hor-mone to stimulate pituitary gonadotropes On the
other hand, sustained administration of the peptide
brings about a short initial stimulation that is rapidly
followed by a decrease in gonadotropin synthesis and
secretion [99] In support of this, the signal response is
different at different doses It has been shown that
pul-satile GnRH stimulates more sustained ERK activity
(more than 8 h), whereas continuous infusion of aT3-1
cells with GnRH stimulates short-term (2 h) ERK
activity [100] There is also evidence that GnRH
treat-ment stimulates cAMP production at nanomolar
con-centrations, but has an inhibitory effect at micromolar
concentrations [101] It should be pointed out that the
nanomolar concentration (0.01–1 nm) corresponds to
the physiological circulating level, and the effects
caused by this dose range may represent the
physiolog-ical functions of GnRH [54,102]
Second, GnRH action has been shown to be
medi-ated by coupling to different Ga-proteins, depending
on the time and dose of exposure [101,103] In general,
Gaq and Gas are associated with a stimulatory effect
[103], whereas Gai often mediates the antiproliferative
and proapoptotic effects of GnRH [73,74] Low GnRH
concentrations promote the coupling of GnRHR to
Gas [101] High concentrations of GnRH cause a
switch in receptor coupling from Gas to Gai [101]
Moreover, stimulation of cAMP production by GnRH
is through coupling to Gas, whereas inhibition of
cAMP production at high concentrations of GnRH is
through coupling to Gai[101,104] These findings
sug-gest that the intracellular milieu in different tissues
results in differential coupling and different phenotypic
effects
Third, multiple GnRH-dependent signaling path-ways may occur via different subunits of a single G-protein [105] After ligand-induced dissociation, both the a-subunit and bc-subunits are capable of activating various effectors, such as adenylyl cyclase, phospholipase C, and ion channels, thereby conferring
on the receptor the potential for dual signaling [106,107] The effector pathway to be activated is specific to the upstream subunits For instance, whereas the a-subunit of Gi inhibits adenylyl cyclase activity, the bc-subunits may stimulate the activities of some adenylyl cyclase subtypes [108,109]
The receptor expression level is also known to be a determinant for different signal outcomes [6,110,111]
In gonadotropes, different cell surface densities of GnRHR result in the differential regulation of luteiniz-ing hormone and follicle-stimulatluteiniz-ing hormone subunit gene expression by GnRH-I [112] We and others have previously shown that low doses of GnRH upregulate the expression of its receptor, whereas high doses decrease it [12,111,113] This difference in regulation suggests that high levels of GnRHR expression may enhance the cellular response to GnRH stimulation, presumably due to more efficient signal amplification
or altered signaling through coupling to different G-proteins
Moreover, ligand selectivity has been proposed to explain the opposite (stimulatory and inhibitory) effects of GnRH For instance, in positively respond-ing prostate carcinoma cell lines, GnRH-I is more effective than GnRH-II On the other hand, in nega-tively responding cell lines, GnRH-II is much more effective than GnRH-I Given the short plasma half-life of GnRH, efforts have been made to obtain GnRH analogs, to resist degradation and to increase potency However, the different GnRH agonists may selectively stabilize different receptor-active conforma-tions and therefore different ligand-induced selective signaling pathways [114] In this regard, it has been shown that the highly variable amino acid at posi-tion 8 of GnRH plays a discriminating role in selecting the receptor conformational state [115]
The presence of splice variants of the GnRHR tran-script may be another possible reason for the different
or opposite responses to GnRH To date, variant tran-scripts of GnRHR have been isolated in many species, e.g chicken [116], mouse [117] and human [118] Although these splice variants are totally incapable of ligand binding or signal transduction, they have been implicated in the functional regulation of the wild-type receptor Previous studies have reported their inhibi-tory activity on full-length GnRHR function [119] This inhibition is specific, augmented by increasing
Table 1 Potential mechanisms that underlie the diverse responses
to GnRH.
Treatment conditions, e.g duration
and dose
[99–101]
Different Gasubtypes [73,74,101,103,104]
Different G-protein subunits [105–109]
Presence of GnRHR splice variants [118,119]
Intrinsic cellular properties [61,62]
Trang 9Original histology
Tumorigenicity in
Anchorage- independent growth
Invasive capability in
Primary carcinoma
Poorly differentiated
Primary carcinoma Moderately differentiated
Primary carcinoma Well differentiated
Lymph node Well differentiated
Pleural effusion
Pleural effusion
Lymph node Poorly differentiated
Trang 10amounts of the cotransfected splice variant cDNA and
possibly by preventing or diverting the normal
process-ing of GnRHR or enhancprocess-ing GnRHR degradation
[118]
Finally, it is also possible that differences in
response may be ascribed to the intrinsic properties of
the cells The physiological characteristics of the
human cancer cell lines mentioned in this minireview
are summarized in Table 2 For example, in contrast
to SKOV-3 cells, OVCAR-3 cells have low invasive
potential Thus, whereas low doses of GnRH-I and
GnRH-II can exert a significant invasive effect in
OVCAR-3 cells, they fail to stimulate SKOV-3
maxi-mally [12,13] Both GnRH-I and GnRH-II only exert
inhibitory effects on SKOV-3 cells at high doses [13]
Novel receptor(s) for GnRH in humans?
An important issue that remains unresolved in this
field is whether one or more other GnRHR subtypes
exist in humans The discovery of GnRH-II has
stimu-lated the search for a cognate type II GnRHR
Molec-ular cloning of the type II GnRHR in goldfish,
marmoset and monkey has shown that the type II
receptor is structurally and functionally distinct from
the type I receptor [138–140] In humans, a type II
GnRHR has not been found However, a search of the
human genome database has revealed a putative
type II GnRHR gene on chromosome 1q21.1
[140,141] Expression of this type II GnRHR mRNA
has been shown in many human tissues, including
endometrium, ovary, placenta, and prostate cancer
cells [42,55,140–142] Although these findings suggest
that the human type II receptor gene is
transcription-ally active, the mRNA is disrupted by a frameshift in
coding exon 1 and a premature stop codon in exon 2,
suggesting that a conventional seven-transmembrane
receptor cannot be translated from this gene The gene
also overlaps two flanking genes and displays
alterna-tive splicing [143] Thus, the functionality of these
human type II GnRHR splice variants and their
involvement in transmitting signals from GnRH-II are
still in question
One noteworthy feature of the primate type II
GnRHR is that, unlike the type I receptor, it possesses
a C-terminal tail, which is responsible for the
recep-tor’s susceptibility to rapid desensitization and
inter-nalization [138,144,145] Finch et al showed that
GnRH was able to efficiently inhibit the proliferation
of breast cancer cells when engineered with sheep
type I GnRHR, but not with Xenopus type II GnRHR
[145] This clearly implies that the antiproliferative
effect of GnRH is mediated most efficiently by a
recep-tor that is not rapidly desensitized or internalized There is evidence that GnRH-II may act through the type I GnRHR In monkey pituitary cultures, in which the type II GnRHR is functional, GnRH-II has been found to stimulate gonadotropin secretion exclusively through the type I GnRHR [146] In contrast, other evidence suggests that the neuromodulatory action of GnRH-II on mammalian behavior is not mediated via the type I receptor in musk shrews [147] Thus, it appears that GnRH-II may selectively interact with different GnRHRs to mediate its different actions, pre-sumably due to the structural differences between the two GnRHR subtypes
Alternatively, it is possible that the human type II GnRHR may be encoded by a different gene that has yet to be identified Database searches have revealed the presence of more than two other GnRHR genes in the human genome apart from the conventional type I receptor gene [148] These genes are located on sepa-rate chromosomes Whether functional, full-length transcripts can be produced from these receptor-like genes remains to be determined Recently, a novel GnRH-II-binding protein, in addition to a conven-tional GnRHR, has been identified by using photo-affinity labeling with an azidobenzoyl-conjugated GnRH-II in prostate cancer cells [149] Taken together, these observations thus suggest the potential existence
of novel receptors for GnRH-I and GnRH-II
Concluding remarks
This overview shows that GnRH modulates a variety
of cellular functions in extrapituitary tissues, such as cell growth, invasion, and angiogenesis However, the effects of GnRH are complex and appear to be cell context dependent The ability of GnRH to elicit very different, even opposite (positive and negative), responses in extrapituitary tissues may arise from dif-ferential usage of signal transduction pathways and receptor cross-talk Clearly, further studies are required to unravel this complex signaling network and the coordinated regulatory roles of different factors in specific cellular events during tumorigenesis High-throughput gene profiling and bioinformatics approaches should be helpful to expand this area of research The information may also serve as a basis for investigators in the field to explore the signaling mechanisms of other G-protein-coupled receptors Most studies thus far have only been conducted in cellular models, but in vivo approaches will be essential for a complete understanding of the specific role of each GnRH isoform, including the putative GnRH-III isolated from the human brain [150] Given the clinical