+ All Categories
Home > Documents > Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone...

Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone...

Date post: 27-Jul-2020
Category:
Upload: others
View: 8 times
Download: 0 times
Share this document with a friend
13
[CANCER RESEARCH 64, 7533–7544, October 15, 2004] Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote Proapoptotic Signaling in Peripheral Reproductive Tumor Cells by Activating a G i -Coupling State of the Type I GnRH Receptor Stuart Maudsley, 1,2 Lindsay Davidson, 1 Adam J. Pawson, 1 Raymond Chan, 1 Rakel Lo ´pez de Maturana, 3 and Robert P. Millar 1,4 1 Medical Research Council Human Reproductive Sciences Unit, Edinburgh, United Kingdom; 2 National Institutes of Health National Institute on Aging, Johns Hopkins Medical Center, Gerontology Research Center, Baltimore, Maryland; 3 Ardana Bioscience, Edinburgh, United Kingdom; and 4 Division of Medical Biochemistry, University of Cape Town Faculty of Health Sciences, Cape Town, South Africa ABSTRACT Gonadotropin-releasing hormone (GnRH) receptor agonists are exten- sively used in the treatment of sex hormone-dependent cancers via the desensitization of pituitary gonadotropes and consequent decrease in steroid sex hormone secretion. However, evidence now points to a direct inhibitory effect of GnRH analogs on cancer cells. These effects appear to be mediated via the G i -type G protein, in contrast to the predominant G q coupling in gonadotropes. Unlike G q coupling, G i coupling of the GnRH receptor can be activated by both agonists and antagonists. This unusual pharmacology suggested that the receptor involved in the cancer cells may not be the classical gonadotrope type I GnRH receptor. How- ever, we have previously shown that a functional type II GnRH receptor is not present in man. In the present study, we show that GnRH agonists and selective GnRH antagonists exert potent antiproliferative effects on JEG-3 choriocarcinoma, benign prostate hyperplasia (BPH-1), and HEK293 cells stably expressing the type I GnRH receptor. This antipro- liferative action occurs through a G i -mediated activation of stress-acti- vated protein kinase pathways, resulting in caspase activation and trans- membrane transfer of phosphatidlyserine to the outer membrane envelope. Structurally related antagonistic GnRH analogs displayed di- vergent antiproliferative efficacies but demonstrated equal efficacies in inhibiting GnRH-induced G q -based signaling. Therefore the ability of GnRH receptor antagonists to exert an antiproliferative effect on repro- ductive tumors may be dependent on ligand-selective activation of the G i -coupled form of the type I GnRH receptor. INTRODUCTION Gonadotropin-releasing hormone (GnRH) is the central regulator of the reproductive hormonal cascade and was first isolated from mam- malian hypothalami (1–3). This decapeptide is synthesized and re- leased by hypothalamic secretory neurones and is delivered to the pituitary gland via the hypophyseal portal blood system. Interaction of the GnRH decapeptide with heptahelical GnRH receptors on pituitary gonadotrope cells induces the release of the pituitary gonadotropin hormones. GnRH has also been found in extrahypothalamic regions of the central nervous system (4) and in nonneuronal tissues such as placenta (5), ovary (6), mammary gland (7), and lymphoid cells (8). The heptahelical type I GnRH receptor is also expressed in these tissues (9). In addition, GnRH I ligand and the type I GnRH receptor are expressed in a number of malignant tumors and cell lines, includ- ing cancers of the breast, ovary, endometrium, and prostate (10). The specific function of GnRH I and its receptor in these extrapituitary sites is unclear. However, an autocrine/paracrine function has been suggested (11, 12). Linked to this hypothesis is the well-documented observation that direct application of GnRH analogs to peripheral reproductive tumor cells results in an attenuation of cellular prolifer- ation and activation of cell death mechanisms (refs. 11–17; for review, see ref. 18). The ability of both GnRH agonists and antagonists to inhibit tumor cell growth suggested that the effects may be mediated by a novel “type II” GnRH receptor, distinct from the cloned pituitary type I receptor at which GnRH agonists stimulate G q and the production of inositol trisphosphate and diacylglycerol that consequently mobilize intracellular calcium and activate protein kinase C. This notion was reinforced by the observation that classical antagonistic GnRH ana- logs induce antiproliferative actions on tumor cells that are mediated by G i activation (for review, see refs. 19 –22). Although a type II GnRH receptor has been cloned from some primates (23, 24), a functional type II GnRH receptor is not present in man (for reviews, see refs. 25 and 26). Moreover, the only functional GnRH receptor transcripts present in human peripheral tissues and tumor cells are identical to type I GnRH receptor sequence expressed in the pituitary (19, 27, 28). Thus, it appears that the difference in cellular milieu between the pituitary gonadotropes and the peripheral sites of type I GnRH receptor expression may be responsible for the differences in GnRH-mediated signaling. The distinctions in GnRH receptor-ligand pharmacology demonstrated in peripheral cells for the activation of G i , as opposed to G q , also suggest that the ligand stabilizes the receptor in a G i -coupling conformation that is different from the conformation mediating G q coupling. The principle of agonist-directed trafficking of receptor signaling [or “ligand-selective signaling,” as we prefer to term it (9, 29)] predicts that when a receptor signals through more than one inde- pendent signal transduction pathway, the relative efficacies/potencies of a series of analogs may differ for the respective pathways (30, 31). This hypothesis builds on the concept that a heptahelical rhodopsin- like G protein-coupled receptor can exist in distinct states (or confor- mations) and that the ability of those states to activate different G protein types may differ. Recently, several examples of such ligand- selective signaling have been demonstrated in biogenic amine of G protein-coupled receptors (32–35). Because there appears to be a specific divergence with respect to the pharmacological profile be- tween the peripheral sites of GnRH action and those in the pituitary, we set out to determine the relationship between the nature of GnRH analogs and their specific effects on differential signal activation. In this study, we have demonstrated that analogs of GnRH that exert a classical antagonist action on cell systems in which the type I GnRH receptor stimulates G q -type mechanisms can differ in their ability to stabilize the specific G i -type G protein-coupling state of the type I GnRH receptor. Thus, we have identified some specific molecular properties of two structurally similar GnRH receptor pep- tide ligands that can determine the capacity of signaling through a specific class of downstream G protein-linked systems. Our discov- eries are particularly pertinent in that the predominant current thera- Received 4/23/04; revised 7/11/04; accepted 8/19/04. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Note: S. Maudsley and L. Davidson contributed equally to this work. Requests for reprints: Robert P. Millar, Medical Research Council Human Repro- ductive Sciences Unit, The University of Edinburgh Chancellor’s Building, 49 Little France Crescent, Edinburgh EH16 4SB, United Kingdom. Fax: 441312426231; E-mail: [email protected]. ©2004 American Association for Cancer Research. 7533 Research. on October 1, 2020. © 2004 American Association for Cancer cancerres.aacrjournals.org Downloaded from
Transcript
Page 1: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

[CANCER RESEARCH 64, 7533–7544, October 15, 2004]

Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ProapoptoticSignaling in Peripheral Reproductive Tumor Cells by Activating aG�i-Coupling State of the Type I GnRH Receptor

Stuart Maudsley,1,2 Lindsay Davidson,1 Adam J. Pawson,1 Raymond Chan,1 Rakel Lopez de Maturana,3 andRobert P. Millar1,4

1Medical Research Council Human Reproductive Sciences Unit, Edinburgh, United Kingdom; 2National Institutes of Health National Institute on Aging, Johns Hopkins MedicalCenter, Gerontology Research Center, Baltimore, Maryland; 3Ardana Bioscience, Edinburgh, United Kingdom; and 4Division of Medical Biochemistry, University of Cape TownFaculty of Health Sciences, Cape Town, South Africa

ABSTRACT

Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent cancers via thedesensitization of pituitary gonadotropes and consequent decrease insteroid sex hormone secretion. However, evidence now points to a directinhibitory effect of GnRH analogs on cancer cells. These effects appear tobe mediated via the G�i-type G protein, in contrast to the predominantG�q coupling in gonadotropes. Unlike G�q coupling, G�i coupling of theGnRH receptor can be activated by both agonists and antagonists. Thisunusual pharmacology suggested that the receptor involved in the cancercells may not be the classical gonadotrope type I GnRH receptor. How-ever, we have previously shown that a functional type II GnRH receptoris not present in man. In the present study, we show that GnRH agonistsand selective GnRH antagonists exert potent antiproliferative effects onJEG-3 choriocarcinoma, benign prostate hyperplasia (BPH-1), andHEK293 cells stably expressing the type I GnRH receptor. This antipro-liferative action occurs through a G�i-mediated activation of stress-acti-vated protein kinase pathways, resulting in caspase activation and trans-membrane transfer of phosphatidlyserine to the outer membraneenvelope. Structurally related antagonistic GnRH analogs displayed di-vergent antiproliferative efficacies but demonstrated equal efficacies ininhibiting GnRH-induced G�q-based signaling. Therefore the ability ofGnRH receptor antagonists to exert an antiproliferative effect on repro-ductive tumors may be dependent on ligand-selective activation of theG�i-coupled form of the type I GnRH receptor.

INTRODUCTION

Gonadotropin-releasing hormone (GnRH) is the central regulator ofthe reproductive hormonal cascade and was first isolated from mam-malian hypothalami (1–3). This decapeptide is synthesized and re-leased by hypothalamic secretory neurones and is delivered to thepituitary gland via the hypophyseal portal blood system. Interaction ofthe GnRH decapeptide with heptahelical GnRH receptors on pituitarygonadotrope cells induces the release of the pituitary gonadotropinhormones. GnRH has also been found in extrahypothalamic regions ofthe central nervous system (4) and in nonneuronal tissues such asplacenta (5), ovary (6), mammary gland (7), and lymphoid cells (8).The heptahelical type I GnRH receptor is also expressed in thesetissues (9). In addition, GnRH I ligand and the type I GnRH receptorare expressed in a number of malignant tumors and cell lines, includ-ing cancers of the breast, ovary, endometrium, and prostate (10). Thespecific function of GnRH I and its receptor in these extrapituitarysites is unclear. However, an autocrine/paracrine function has been

suggested (11, 12). Linked to this hypothesis is the well-documentedobservation that direct application of GnRH analogs to peripheralreproductive tumor cells results in an attenuation of cellular prolifer-ation and activation of cell death mechanisms (refs. 11–17; for review,see ref. 18).

The ability of both GnRH agonists and antagonists to inhibit tumorcell growth suggested that the effects may be mediated by a novel“type II” GnRH receptor, distinct from the cloned pituitary type Ireceptor at which GnRH agonists stimulate G�q and the production ofinositol trisphosphate and diacylglycerol that consequently mobilizeintracellular calcium and activate protein kinase C. This notion wasreinforced by the observation that classical antagonistic GnRH ana-logs induce antiproliferative actions on tumor cells that are mediatedby G�i activation (for review, see refs. 19–22). Although a type IIGnRH receptor has been cloned from some primates (23, 24), afunctional type II GnRH receptor is not present in man (for reviews,see refs. 25 and 26). Moreover, the only functional GnRH receptortranscripts present in human peripheral tissues and tumor cells areidentical to type I GnRH receptor sequence expressed in the pituitary(19, 27, 28). Thus, it appears that the difference in cellular milieubetween the pituitary gonadotropes and the peripheral sites of type IGnRH receptor expression may be responsible for the differences inGnRH-mediated signaling. The distinctions in GnRH receptor-ligandpharmacology demonstrated in peripheral cells for the activation ofG�i, as opposed to G�q, also suggest that the ligand stabilizes thereceptor in a G�i-coupling conformation that is different from theconformation mediating G�q coupling.

The principle of agonist-directed trafficking of receptor signaling[or “ligand-selective signaling,” as we prefer to term it (9, 29)]predicts that when a receptor signals through more than one inde-pendent signal transduction pathway, the relative efficacies/potenciesof a series of analogs may differ for the respective pathways (30, 31).This hypothesis builds on the concept that a heptahelical rhodopsin-like G protein-coupled receptor can exist in distinct states (or confor-mations) and that the ability of those states to activate different Gprotein types may differ. Recently, several examples of such ligand-selective signaling have been demonstrated in biogenic amine of Gprotein-coupled receptors (32–35). Because there appears to be aspecific divergence with respect to the pharmacological profile be-tween the peripheral sites of GnRH action and those in the pituitary,we set out to determine the relationship between the nature of GnRHanalogs and their specific effects on differential signal activation.

In this study, we have demonstrated that analogs of GnRH thatexert a classical antagonist action on cell systems in which the type IGnRH receptor stimulates G�q-type mechanisms can differ in theirability to stabilize the specific G�i-type G protein-coupling state ofthe type I GnRH receptor. Thus, we have identified some specificmolecular properties of two structurally similar GnRH receptor pep-tide ligands that can determine the capacity of signaling through aspecific class of downstream G protein-linked systems. Our discov-eries are particularly pertinent in that the predominant current thera-

Received 4/23/04; revised 7/11/04; accepted 8/19/04.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Note: S. Maudsley and L. Davidson contributed equally to this work.Requests for reprints: Robert P. Millar, Medical Research Council Human Repro-

ductive Sciences Unit, The University of Edinburgh Chancellor’s Building, 49 LittleFrance Crescent, Edinburgh EH16 4SB, United Kingdom. Fax: 441312426231; E-mail:[email protected].

©2004 American Association for Cancer Research.

7533

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 2: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

pies for hormone-dependent cancers, such as prostate cancer, useGnRH analogs that inhibit sex hormone production. Although thistherapy provides amelioration of the disease in the short term, this isoften followed by aggressive recurrence in the form of sex hormone-independent cancers. Thus, the identification of GnRH analogs withdirect effects on cancer cells offers the opportunity of targeting themin conjunction with, or independently from, hormone depletion.

MATERIALS AND METHODS

Reagents. The p38/c-Jun NH2-terminal kinase (JNK) inhibitor SB203580and the MAPK-kinase 1(2) (MEK1/2) inhibitor PD98059 were obtained fromCalbiochem (La Jolla, CA) and prepared in dimethyl sulfoxide (DMSO; finalDMSO concentration, 0.1% in cell treatments). Fluorescein isothiocyanate(FITC)-conjugated annexin V and anti-myc sera were obtained from SantaCruz Biotechnology (Santa Cruz, CA). Anti-cleaved and pro-caspase-3, anti-active extracellular signal-regulated kinase (ERK) 1/2, ERK5, JNK, and p38sera were all obtained from New England Biolabs (Beverly, MA). Pertussistoxin (PTX) was obtained from Affiniti Research Products (Biomol, PlymouthMeeting, PA). GnRH I/II, lysophosphatidic acid (LPA), and forskolin (FSK)were obtained from Sigma (St. Louis, MO). Antagonists 135-18, 135-25, 6,and 21 were generously supplied by Roger Roeske (University of Indiana,Indianapolis, IN). The myc-tagged mitogen-activated protein kinase (MAPK)cDNA isoforms JNK and p38� were generously supplied by Eisuke Nishida(Kyoto University, Kyoto, Japan).

Cell Culture and Transfection. Human benign prostate hyperplasia(BPH-1), human JEG-3 choriocarcinoma (American Type Culture Collection,Manassas, VA), HEK293 cells stably expressing the type I GnRH receptor(designated as SCL60; ref. 36) and �T4-gonadotropes (obtained from PamelaMellon; University of California at San Diego, San Diego, CA) stably express-ing the marmoset type II GnRH receptor (designated as �T4-II; ref. 23) weremaintained in Dulbecco’s modified Eagle’s medium (Sigma) supplementedwith 10% fetal bovine serum, 2% glutamine, and 1% penicillin (10,000units/mL)/streptomycin (10,000 �g/mL) at 37°C in a humidified 5% CO2

atmosphere. Where required, cells were serum-deprived by incubation for 16hours in Dulbecco’s modified Eagle’s medium supplemented with only glu-tamine and penicillin/streptomycin. Ligands were applied to cells at 37°C forthe time periods specified in figure legends. Chemical inhibitors were prein-cubated with the cells before agonist stimulation for the time periods specifiedin the figure legends. Transient transfections of JEG-3 or BPH-1 cells wereperformed using Superfect (Qiagen, Valencia, CA) according to the manufac-turer’s instructions.

Immunoprecipitation and Immunoblotting. After stimulation, cytoplas-mic proteins were extracted as described previously (37). Proteins were re-solved by SDS-PAGE for confirmation of plasmid expression or determinationof intracellular protein activation by immunoblotting. Immunoprecipitation ofmyc epitope-tagged proteins was achieved by addition of 25 �L of a 30%slurry of anti-myc agarose preconjugated antisera to the clarified cell lysate(Santa Cruz Biotechnology) with agitation for 16 hours at 4°C. Immunocom-plexes were collected by centrifugation (10,000 � g, 10 minutes) and washedtwice in ice-cold Nonidet P-40–based solubilization buffer (37) before addi-tion of 25 �L of Laemmli sample buffer. Immunoprecipitates were resolved bySDS-PAGE and transferred to polyvinylidene difluoride membrane (NEN LifeSciences, Boston, MA) for protein immunoblotting. Polyvinylidene difluoridemembranes were blocked in a 4% bovine serum albumin, 50 mmol/L Tris-HCl(pH 7.0), 0.05% Tween 20, and 0.05% Nonidet P-40 blocking solution.

Immunoblotting of the active forms of ERK1/2, p38, and JNK was per-formed as described previously by Millar et al. (23). Phosphorylation ofERK1/2, JNK, p38, or ERK5 was detected with a 1:1,000 dilution of anti–phospho-specific ERK1/2, JNK, p38, or ERK5 rabbit polyclonal antibodies,respectively (New England Biolabs). The extent of MAPK activation wasassessed and normalized by subsequently applying antisera (1:1,000 dilution)against the unphosphorylated forms of ERK2, JNK, or p38 (New EnglandBiolabs) to primary antibody-stripped immunoblots. An alkaline phosphatase-conjugated IgG (Sigma) was used as a secondary antibody for anti-activeERK1/2/JNK/p38 and unphosphorylated ERK1/2/JNK/p38. Visualization ofalkaline phosphatase-labeled proteins was performed using enzyme-linkedchemifluorescence Amersham Pharmacia Biotech (Piscataway, NJ) and quan-

tified using a Molecular Dynamics (Sunnyvale, CA) Storm 860 PhosphorIm-ager.

Cell Proliferation. The proliferation of SCL60, BPH-1, and JEG-3 cellswas measured by counting the number of viable trypan blue-excluding cellsafter 5 days of continuous GnRH receptor-interacting ligand exposure. Cellswere plated at an initial minimal confluence (10–20%) to allow for 5 days ofcontinual growth that would not result in 100% cell confluence by day 5. Cellswere replenished with new ligand every 12 hours, and the number of viablecells compared with vehicle-treated control cells was measured. AnnexinV-FITC staining was performed on cells treated for various time periods withGnRH by immersing live cells in starving media supplemented with a 1:100dilution of annexin V-FITC for 30 minutes before fixing the cells in 100%

Fig. 1. Antiproliferative effects of GnRH receptor-interacting ligands on humanchoriocarcinoma cells (JEG-3), human benign prostate hyperplastic cells (BPH-1), andtype I GnRH receptor-expressing human embryonic kidney cells (SCL60). JEG-3 (A),BPH-1 (B), and SCL60 (C) cells were treated continuously for 5 days with the indicateddoses of GnRH I (f), GnRH II (Œ), antagonist 135-18 (�), or antagonist 135-25 (�)added directly to the growth medium. Ligands were replenished every 24 hours. Viablecells were counted in triplicate at the end of the 5-day period. The data represent themean � SE of four experiments.

7534

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 3: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

methanol for 10 minutes at �20°C and mounting them in Permafluor (Immu-notech, Marseilles, France). Annexin V-FITC–reactive cells were observedusing a Zeiss LSM510 confocal scanning laser microscope. The GnRH-induced expression of cleaved and pro-caspase-3 was assessed by specificimmunoblot. The expression of caspases was measured by immunoblotting forthe specific cleaved or pro-caspase forms with rabbit polyclonal antisera(1:1,000 dilutions; New England Biolabs) using an antirabbit alkaline phos-phatase-conjugated sera (1:10,000 dilution) as a secondary antibody. Proteinswere detected by enzyme-linked chemifluorescence measured with a Molec-ular Dynamics Storm 860 PhosphorImager.

Phosphatidylinositol Hydrolysis. Inositol phosphate production was as-sayed as described previously by prelabeling cells with myo-[3H]inositol(Amersham Pharmacia Biotech) and measuring [3H]inositol phosphates afterGnRH stimulation (23, 24).

Intracellular Cyclic AMP Measurement. Intracellular cyclic AMP(cAMP) concentration in BPH-1 or JEG-3 cells was measured using a propri-etary fluorescent cAMP assay kit according to the manufacturer’s instructions(Biomol). GnRH pretreatments (for the time periods specified in the appro-priate figure legends) were made before the standard 15-minute FSK applica-tion to the cells (1 �mol/L, JEG-3 cells; 3 �mol/L, BPH-1 cells).

RESULTS

GnRH and GnRH Analogs Mediate Antiproliferative Effects onJEG-3, BPH-1, and HEK293 Cells. Continuous treatment of JEG-3human choriocarcinoma with GnRH I, GnRH II, and the antagonist135-25 [Ac-D-Nal(2)-D-4-ClPhe-D-Pal-Ser-1-MePal-D-IsopropylLys-Leu-IsopropylLys-Pro-D-AlaNH2] resulted in a dose-response inhibitionof cellular growth. Antagonist 135-18 [Ac-D-Nal(2)-D-4-ClPhe-D-Pal-Ser-Ile-D-IsopropylLys-Leu-IsopropylLys-Pro-D-AlaNH2], despite a sin-

gle amino acid difference in position 5 of the decapeptide, failed todemonstrate an antiproliferative effect of a magnitude similar to thatgenerated by antagonist 135-25, GnRH I, and GnRH II (Fig. 1A). Thehyperplastic prostate cell line BPH-1 exhibited a similar profile of anti-proliferative response to the same panel of GnRH analogs (Fig. 1B).

We further investigated the nature of the ligand-receptor specificityof the antiproliferative effects of GnRH receptor systems by studyingthe effects of the same panel of ligands on a model cell background,i.e., SCL60 cells. After continuous treatment of the SCL60 cells withGnRH I (Fig. 1C), there was a dramatic reduction in the growth of thecells and a substantial decrease in total cell number at the end of theexperimental period. As with the JEG-3 and BPH-1 cells, GnRH IIdemonstrated a capacity similar to that of GnRH I when inhibiting cellgrowth. The two GnRH receptor antagonists behaved in a similarmanner as in the JEG-3 and BPH-1 cells, in that antagonist 135-18proved to be relatively ineffective at inhibiting the SCL60 cell pro-liferation, whereas antagonist 135-25 was nearly as effective as GnRHI and GnRH II. Compared with either JEG-3 or BPH-1 cells, there wasconsiderably greater inhibition of cell growth and a greater degree ofdetectable cell death apparent from 48 hours onward of GnRH stim-ulation. We attribute this greater effect of the GnRH receptor ligandson SCL60 cells to the much greater level of receptor expression inthese cells. Only minimal levels of cell surface receptor expressionwere noted in JEG-3 and BPH-1 cells (not in excess of 200 specificcpm for 125I-His5-D-Tyr6-GnRH I; data not shown), whereas as muchas a 10- to 20-fold greater expression level was demonstrated inSCL60 cells.

Recent reports have suggested that the atypical GnRH receptor

Fig. 2. Pharmacological profile suggests that type I GnRH receptor mediates a peripheral antiproliferative effect. Using two cell lines expressing a single GnRH receptor population,it can be demonstrated that antagonist 135-25 does not seem to exert any significant biological activity at the type II GnRH receptor, which is proposed to be the locus of GnRH actionin peripheral tumor cells. The histograms in A and B represent inositol phosphate accumulation data from SCL60 cells (which express the type I GnRH receptor). In A, GnRH I (10nmol/L) exerts typical agonistic activity, whereas antagonist 135-18 appears to have little agonistic capacity. Antagonist 135-25 displays no agonistic activity compared with GnRHI. In C and D, inositol phosphates liberated were measured in the gonadotrope cells (�T4) stably expressing the marmoset type II GnRH receptor (designated �T4-II). In C, stimulationwith GnRH II (10 nmol/L) induces significant inositol phosphate accumulation. Antagonist 135-18 clearly exerts a dose-dependent agonistic effect on the type II GnRHreceptor-expressing cells. D. Antagonist 135-25 fails to liberate any free inositol phosphates.

7535

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 4: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

pharmacology observed in peripheral reproductive tumor lines is dueto the expression of a type II human GnRH receptor, similar to thatoriginally cloned by Millar et al. (23). However, the data abovesuggest that it is unlikely that the antiproliferative GnRH effectsobserved occur via a type II GnRH receptor because antagonist135-18 possess a high degree of agonistic activity on type II GnRHreceptors cloned from several species (23, 38). In the presence a typeII receptor, antagonist 135-18 would be expected to exert a profoundantiproliferative effect; however, it had the lowest effective antipro-liferative capacity observed among the four GnRH analogs. More-over, a full-length type II receptor cannot be transcribed from thehuman gene due to a frameshift and premature stop codon (39).

Antagonist 135-25 Demonstrates a Selective Type I Gonado-tropin-Releasing Hormone Receptor Activation Profile. We havedemonstrated that despite a high degree of similarity between antag-onist 135-18 and antagonist 135-25, there is a significant difference intheir capacity to stimulate certain forms of GnRH receptor activity,i.e., an antiproliferative action on both tumorous and hyperplastic celllines. When compared in SCL60 cells expressing a type I GnRHreceptor, both antagonists 135-18 and 135-25 demonstrate no agonis-tic activity (Fig. 2A and B). Both agents can also efficiently inhibitGnRH-mediated accumulation of inositol phosphates at the type IGnRH receptor (data not shown). When the two antagonists arecompared in a cellular background solely expressing the marmosettype II GnRH receptor, i.e., �T4-II cells (23), antagonist 135-18

demonstrated considerable agonistic activity (Fig. 2C), whereas an-tagonist 135-25 showed no activity at all (Fig. 2D). If the antiprolif-erative actions of GnRH and related ligands on the cell types used inthis study (BPH-1 and JEG-3) were mediated through a type II-likehuman GnRH receptor, then antagonist 135-18 may be expected topossess a greater antiproliferative capacity than antagonist 135-25. Itis therefore extremely unlikely, from both a pharmacological andmolecular biological viewpoint, that the antiproliferative actions ofGnRH I or antagonist 135-25 are occurring via stimulation of a humanType II-like GnRH receptor.

Gonadotropin-Releasing Hormone Activates G�i-Mediated Re-ceptor Signaling Pathways in JEG-3 and BPH-1 Cells. It has beenshown in several reports that the G protein coupling of GnRH recep-tors expressed in peripheral tumor tissues differs from that of theanterior pituitary. In the pituitary, the primary G protein couplingevent of the stimulated GnRH receptor is via the G�q-type G protein.In contrast, in peripheral tissues, the primary GnRH receptor Gprotein coupling event appears to be via the PTX-sensitive G�i-typeG protein pathway. We tested whether GnRH induced the activationof G�i-mediated signaling pathways in our experimental paradigms.FSK stimulation of JEG-3 (1 �mol/L) and BPH-1 (3 �mol/L) cells(inducing a 50% Rmax cAMP response in each case) was blunted withextended cellular pretreatment times (10–60 minutes) with either 100nmol/L GnRH I or antagonist 135-25 [Fig. 3A (JEG-3) and B (BPH-1)]. In Fig. 3C (JEG-3) and D (BPH-1), the ability of the 60-minute

Fig. 3. GnRH I and antagonist 135-25 activatethe G�i-type G protein. In A (JEG-3) and B (BPH-1), a single 15-minute stimulation with FSK (1�mol/L, JEG-3; 3 �mol/L, BPH-1) potently ele-vates intracellular cAMP levels. With increasingtimes of preexposure (10, 30, and 60 minutes) toGnRH I (100 nmol/L) or antagonist 135-25 (100nmol/L), the FSK-stimulated cAMP accumulationwas significantly reduced. In C (JEG-3) and D(BPH-1), the GnRH I- and antagonist 135-25-me-diated inhibition of FSK-stimulated cAMP accu-mulation (60-minute pre-exposure) is inhibited bypreincubation of the cells with PTX (200 ng/mL, 16hours). E (JEG-3) and F (BPH-1) demonstrate dif-ferential ability of GnRH I, antagonist 135-25, andantagonist 135-18 preincubation (100 nmol/L, 60minutes) to attenuate FSK-induced cAMP accumu-lation. Each histogram in A�F depicts themean � SE of three to four experimental replicatesof the cAMP accumulation assay.

7536

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 5: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

GnRH I or antagonist 135-25 pretreatments to inhibit the FSK-mediated cAMP accumulation was attenuated by a 16-hour pretreat-ment with 200 ng/mL PTX. It therefore appears that both ligands canefficiently activate the adenylate cyclase inhibitory activity of G�i inboth cell models tested. However compared with antagonist 135-25and GnRH I, antagonist 135-18 was significantly less able to inhibitFSK-induced cAMP accumulation in both JEG-3 (Fig. 3E) and BPH-1cells (Fig. 3F).

The GnRH activation of MAPKs in peripheral tumor cells has beenreported to be accompanied by a lack of inositol phosphate turnover.Our studies also demonstrated that there was no appreciable inositolphosphate turnover, even at high doses (0.1–1 �mol/L) of GnRH andits analogs. However, at even higher doses (10–50 �mol/L), GnRH I,GnRH II, and antagonist 135-25 all displayed a small capacity toinduce inositol phosphate accumulation [Fig. 4A�D (JEG-3) and F�I(BPH-1)]. The level of doses required to accomplish this suggests thatactivation of inositol turnover was being mediated by G�� subunits ofanother G protein, e.g., G�i rather than G�q activation of phospho-lipase C-�. In Fig. 4I (JEG-3) and J (BPH-1), we demonstrated thatthe minimal inositol phosphate turnover induced by GnRH was sen-sitive to pretreatment with PTX, whereas the more robust inositolphosphate turnover induced by LPA treatment (activating the endog-enous G�q-coupled endothelial differentiation gene receptor) wascompletely insensitive to the PTX. Thus it appears that there is

negligible inositol phosphate turnover induced in peripheral tumorcells by submicromolar concentrations of GnRH. However, at muchhigher doses, there is a PTX-sensitive capacity to minimally stimulateinositol phosphate turnover, presumably via the Gi��-mediated acti-vation of phospholipase C-�. Exposure of either JEG-3 or BPH-1 cellsto 1 �mol/L LPA, potently stimulating ERK1/2 via the endogenousPTX-insensitive G�q signaling pathway, failed to significantly atten-uate cell growth. In contrast to GnRH receptor activation, an elevationin cell number after 5 days of continuous LPA treatment occurred(data not shown). Thus, it appears that the downstream effects ofERK-mediated pathways are not directly linked to the eventual anti-proliferative effects.

Gonadotropin-Releasing Hormone Activates Stress-ActivatedProtein Kinase Pathways in JEG-3 and BPH-1 Cells. Severalgroups have demonstrated that GnRH treatment of tumor cell linesinduces a potent stimulation of the ERK isoforms of the MAPKfamily (40, 41). We observed a protracted activation of ERK1/2 inJEG-3 cells (Fig. 5A) and BPH-1 cells (Fig. 5B). We additionallyassessed whether GnRH stimulation of either JEG-3 or BPH-1 cellsresulted in significant activation of any other of the MAPK isoforms.We observed no GnRH-specific activation of ERK5/Big-MAPK iso-form of MAPK (data not shown). However, there was a distinctGnRH-induced activation of JNK in JEG-3 cells and a GnRH-inducedactivation of p38 in BPH-1 cells. It therefore appeared that there was

Fig. 4. Induction of inositol phosphate accumulation by GnRH receptor-interacting ligands. A (GnRH I), B (GnRH II), C (Ant 135-18), and D (Ant 135-25) demonstrate that selectiveGnRH receptor-interacting ligands can mediate, in a dose-dependent manner, minimal inositol phosphate accumulation in JEG-3 cells. E, GnRH I activates minimal inositol phosphateaccumulation in a PTX (200 ng/mL, 16-hour pretreatment)-sensitive manner (JEG-3 cells). LPA-induced inositol phosphate accumulation is greater in degree but is PTX insensitive.F (GnRH I), G (GnRH II), H (antagonist 135-18), and I (antagonist 135-25) demonstrate minimal ligand-induced inositol phosphate accumulation in BPH-1 cells. J, GnRH I-induced(but not LPA-induced) accumulation of inositol phosphates is sensitive to PTX preincubation (200 ng/mL, 16 hours, BPH-1 cells). The data in each histogram represent the mean � SEfrom at least three independent experiments.

7537

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 6: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

a degree of cell specificity of MAPK stimulation with GnRH. Tofurther investigate these stress-activated protein kinase (SAPK) acti-vation events, we transfected the tumor cell lines with myc-taggedJNK2 or p38� MAPK isoforms. GnRH caused a time-dependent andprotracted activation of the immunoprecipitated JNK2 in JEG-3 cells(Fig. 5C), and p38� in BPH-1 cells (Fig. 5D) that was considerablydelayed (30 minutes) in contrast to ERK activation (20 minutes).Recent data have demonstrated that an inhibitory effect on epidermalgrowth factor receptor activity may underlie the antiproliferativeaction of GnRH analogs (19). However no demonstrable GnRH-induced reduction in the phosphorylation status of the epidermalgrowth factor receptor in either JEG-3 or BPH-1 cells was observed(data not shown).

We have demonstrated that continuous stimulation with eitherGnRH I or antagonist 135-25 retarded the growth of both JEG-3 andBPH-1 cells (Fig. 1). However, when these experiments were per-formed using antagonist 135-18, which differs from antagonist 135-25by only one amino acid, the antiproliferative effect was minimal (Fig.1). We investigated whether this phenomenon of low antiproliferativepotency of antagonist 135-18 resided in its capacity (or incapacity) toactivate the G�i-SAPK pathways. GnRH I, GnRH II, and antagonist135-25 all activated ERK1/2 to a substantially greater extent thanantagonist 135-18 (all 1 �M, 10 minutes) in JEG-3 cells (Fig. 6A) andBPH-1 cells (Fig. 6B). When tested for its capacity to activate eitherJNK2 in JEG-3 cells or p38� in BPH-1 cells, antagonist 135-18

demonstrated a dramatically lower efficacy than GnRH I or antagonist135-25 with respect to the activation of SAPK isoforms [Fig. 6C(JEG-3) and D (BPH-1)].

By using transfected myc-JNK in JEG-3 cells or myc-p38� inBPH-1 cells, we investigated whether the previously demonstratedGnRH/antagonist 135-25-induced SAPK activation was mediatedthrough a G�i-dependent mechanism. As demonstrated in Fig. 7,treatment with either GnRH I (100 nmol/L, 40 minutes) or antagonist135-25 (100 nmol/L, 40 minutes) led to JNK or p38� activation inJEG-3 (Fig. 7A) or BPH-1 cells (Fig. 7D), respectively, which wasinhibited by PTX. This suggests that both ligands activate SAPKsthrough G�i-type G protein pathways (JEG-3, Fig. 7B and C; BPH-1,Fig. 7E and F).

Antagonist 135-25 but not Antagonist 135-18 Potently Stimu-lates the Activation of G�i–Stress-Activated Protein Kinase Path-ways. When we screened a panel of classical GnRH receptor G�q

antagonists for their ability to stimulate the SAPK pathway and inhibitBPH-1 cell growth, we noted that their ability to stimulate p38 activation(Fig. 8A) was related to their capacity to inhibit BPH-1 cell proliferation(Fig. 8B). It therefore appears that antagonist 135-25, like GnRH I, canadequately activate the G�i-type pathway in JEG-3 or BPH-1 cells,whereas the chemically related antagonist 135-18 has a much lowerpotency with respect to this form of atypical GnRH receptor activation.We suggest that this inability of antagonist 135-18 (and other antagonists)to induce a productive coupling between the GnRH receptor and the

Fig. 5. GnRH I stimulates ERK1/2, JNK, and p38 mitogen-activated protein kinases in both JEG-3 and BPH-1 cells. A and B, time course of GnRH I (100 nmol/L)-induced ERK1/2activation in JEG-3 and BPH-1 cells, respectively. Each bar of the histograms in A and B represents the mean � SE. Data represent three to four experimental replicates. C, GnRH(100 nmol/L)-induced activation of immunoprecipitate myc-tagged JNK in JEG-3 cells is protracted in nature (up to 120 minutes). D, GnRH (100 nmol/L)-induced activation ofimmunoprecipitated myc-tagged p38 in BPH-1 cells is protracted in nature (up to 120 minutes). The respective histograms in C and D represent the mean � SE. Data represent themean of three experimental replicate time courses.

7538

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 7: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

G�i-type G protein pathway in these model cells may account for its poorantiproliferative efficacy.

Gonadotropin-Releasing Hormone Induces the Generation ofProapoptotic States in JEG-3 and BPH-1 Cells. To determinewhether GnRH induced apoptosis, we measured the effects of GnRH onthe structural integrity of the cells’ plasma membrane. A well-docu-mented early event in apoptosis is the reversal in polarity of plasmamembrane constituents such as phosphatidylserine [PS (42)]. UsingFITC-conjugated annexin V protein, which has a high affinity for ex-posed PS, we tested whether exposure of JEG-3 or BPH-1 cells to GnRHresulted in the expression of PS on the membrane outer envelope. After24 hours of subculture of JEG-3 cells in the absence or presence of GnRH(100 nmol/L), no annexin FITC binding could be demonstrated (Fig. 9A,1–6), but after 48 hours of GnRH exposure, a considerable amount ofexternal membrane annexin V-reactive PS was evident (Fig. 9A, 10–12).Cells not exposed to GnRH failed to exhibit any external membraneannexin V-FITC staining even after 48 hours (Fig. 9A, 7–9). Similarresults were obtained from BPH-1 cells for the same period of GnRH Istimulation (Fig. 9A, 13–15). In addition to the generation of early plasmamembrane PS reversal, we demonstrated the generation of proapoptotic

caspase enzymes that have been shown to be involved in cell degradationin many tissues (for review, see ref. 43). In JEG-3 cells continuouslyexposed to GnRH I (100 nmol/L, 24–96 hours), there was a substantialelevation in the cellular levels of pro-caspase 3 (Fig. 9B and D). GnRHI elevation of active caspase-3 levels took longer to emerge and was onlysubstantially evident between 48 and 72 hours of incubation (Fig. 9B andC). A similar pattern of time-dependent increases in pro-caspase andcleaved caspase-3 was evident in BPH-1 cells after GnRH I treatment(Fig. 9E�G).

Gonadotropin-Releasing Hormone-Induced Activation of Stress-Activated Protein Kinase Pathways Is Involved in the Induction of aProapoptotic State. We investigated whether there was a connectionbetween the capacity of GnRH to activate the SAPK pathways and theobserved generation of the early signs of apoptosis, e.g., the PS transferfrom the internal face of the plasma membrane envelope to the externalface. To this end, we used the SAPK inhibitor SB203580, which at lowdoses (1 �mol/L) acts as a potent inhibitor of p38 SAPK activity and athigher doses (20 �mol/L) exerts an additional inhibitory activity on theJNK family of SAPK proteins (44). Coincubation of JEG-3 cells with 20�mol/L SB203580 and GnRH I (100 nmol/L) for 48 hours resulted in a

Fig. 6. GnRH receptor-interacting ligands stimulate ERK1/2, JNK, and p38 mitogen-activated protein kinases in both JEG-3 and BPH-1 cells. A and B, ERK1/2 activation by severalGnRH receptor-interacting ligands (all at 100 nmol/L for 10 minutes) in JEG-3 and BPH-1 cells, respectively. GnRH I and antagonist 135-25 activate ERK1/2 to a similar extent, withGnRH II being less effective, and antagonist 135-18 being the least effective. Each bar of the histograms in A and B represents the mean � SE. Data represent the mean of three tofour experimental replicates. C, GnRH I and antagonist 135-25 (both at 100 nmol/L, 40 minutes) activate myc-JNK in JEG-3 cells, whereas antagonist 135-18 (100 nmol/L, 40 minutes)fails to activate JNK. The immunoblots at the top depict the increasing phosphorylation status of the immunoprecipitated JNK, whereas the level of total unphosphorylated protein(detected with �-JNK antisera) remains unchanged. The histogram at the bottom depicts the mean � SE. Data represent the mean of three experimental replicates of the above Westernblot. D. In BPH-1 cells, both GnRH I and antagonist 135-25 effectively activate p38, whereas antagonist 135-18 does not. The immunoblots at the top depict an increasingphosphorylation of p38 with no increase in total p38 protein. The histogram at the bottom depicts the mean � SE. Data represent the mean of three experimental replicates of the aboveWestern blot.

7539

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 8: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

significant reduction in the degree of annexin V-FITC staining of theexternal aspect of the plasma membrane (Fig. 10A, compare 5 with 8).There was no significant difference in the general growth patterns andgross morphology of the cells treated with SB203580 compared withthose treated with GnRH alone or unstimulated cells (data not shown).Similarly in BPH-1 cells treated with SB203580 (1 �mol/L) for 48 hours,the GnRH-mediated induction of annexin V-FITC reactivity on the outerplasma membrane envelope was almost completely abolished (Fig. 10B,8). As in the JEG-3 cell experiment, there was no significant observablechange in cell morphology or growth rates in either the SB203580-treatedcells or the DMSO vehicle-treated cells. In addition, we demonstratedthat the GnRH-induced generation of annexin V-FITC–reactive cells wasunaffected by continuous treatment with 10 �mol/L PD98059 (an inhib-

itor of MEK1/2), which completely inhibited the capacity of all of theGnRH analogs to activate ERK1/2 in both cell types (data not shown).Thus, it appears that inhibition of the GnRH-induced SAPK pathways inJEG-3 cells and BPH-1 cells can attenuate the capacity of GnRH toinduce apoptotic signs in these two cell lines. Because SB203580 is nothighly specific, confirmatory studies involving the generation of stableJEG-3 and BPH-1 cell lines expressing dominant negative JNK and p38or small interfering RNAs are needed.

DISCUSSION

In this study we have demonstrated that some classical pituitary,G�q signal-inhibiting, GnRH receptor antagonists can act in the same

Fig. 7. GnRH I and antagonist 135-25 activation of JNK or p38 is dependent on G�i stimulation. A depicts a representative Western blot of immunoprecipitated myc-JNK fromJEG-3 cells stimulated for 40 minutes with either GnRH I (100 nmol/L) or antagonist 135-25 (100 nmol/L). The GnRH I and antagonist 135-25–induced increases in JNKphosphorylation are abolished with PTX preincubation (200 ng/mL, 16 hours). The histograms in B and C depict the mean � SE. Data represent the mean of three experimentalreplicates of the above Western blot experiments. D depicts a representative Western blot of immunoprecipitated myc-p38 from BPH-1 cells stimulated for 40 minutes with either GnRHI (100 nmol/L) or antagonist 135-25 (100 nmol/L). The GnRH I and antagonist 135-25–induced increases in p38 phosphorylation are abolished with PTX preincubation (200 ng/mL,16 hours). The histograms in E and F depict the mean � SE. Data represent the mean of three experimental replicates of the above Western blot experiments.

Fig. 8. Ligand-specific antiproliferative effects in BPH-1 cells. A. Diverse GnRH receptor antagonists possess differential ability to stimulate p38 in BPH-1 cells. The antagonistsused (100 nmol/L, 40 minutes) were 6 (Ac-D-Nal (2)-D-Me-4-ClPhe-D-Trp-Ser-Tyr-D-Arg-Leu-Arg-Pro-D-AlaNH2), 21 (Ac-D-Nal(2)-D-4-ClPhe-D-Trp-Ser-His-D-Arg-Pro-D-AlaNH2),and cetrorelix. Data shown are the mean � SE and represent the mean from three individual experiments. B. After 5 days of continuous treatment with each agent at the dose specifiedon the histogram, a differential effect between the various antagonist peptides was observed. Data shown are the mean � SE from three individual experiments.

7540

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 9: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

manner as classical agonists in inhibiting tumor cell proliferation. Theability of these antagonists to exert antiproliferative effects on tumorcells is related to the altered pharmacological profile of GnRH recep-tor signaling in these cells. The differences in efficacy betweenantagonist 135-25 and antagonist 135-18 with respect to their anti-proliferative action appear to be due to the relative abilities of the twopeptides to stabilize an active form of the receptor that is capable ofcoupling productively to G�i. Therefore, the designation of antagonist135-25 as a GnRH receptor “antagonist” is somewhat spurious be-cause this only describes its ability to stabilize the G�q-preferringform/state of the type I GnRH receptor, which is the predominantsignaling mechanism in pituitary gonadotrophs. Hence we have dem-onstrated that antagonist 135-25 exerts a potent antiproliferative ac-tion in JEG-3 and BPH-1 cells due to its ability to activate a G�i-coupling form of the type I GnRH receptor while being unable toinduce GnRH receptor G�q coupling. We have demonstrated that thetwo endogenous forms of GnRH (GnRH I and GnRH II) and alsosome classical antagonists can attenuate cell growth (Fig. 1) and that

this antiproliferative effect is coincident with the induction of aproapoptotic state of the cells, as demonstrated physically by PSmembrane translocation and the activation of proteolytic caspases(Figs. 8 and 9). Activation of the GnRH receptor appears to mediatethese effects via a potent stimulation of members of the SAPKpathway, i.e., JNK and p38 (Figs. 5 and 6), through a G�i-dependentmechanism (Figs. 3, 4, and 9). Thus, certain GnRH analogs are ableto stabilize a specific active conformation of the GnRH receptor thatwill potently convey an antiproliferative effect in peripheral tissues,but not in the pituitary. Whereas the studies here have focused on theapoptotic pathway associated with the G�i-mediated antiproliferativeeffects of GnRHs and analogs, we have also shown an inhibition ofthymidine incorporation and cell cycle arrest (data not shown) as hasbeen described previously (45). The relative contributions of thesemechanisms to the net antiproliferative effects of GnRH and analogshave not been determined.

Initial hypotheses concerning the nature of the divergence in sig-naling between GnRH-responsive sites in the pituitary and those in

Fig. 9. GnRH-induced apoptotic events in JEG-3 and BPH-1 cells. A, GnRH-induced plasma membrane translocation of PS measured using annexin V-FITC recombinant protein.All 12 panels depict either phase-contrast (1, 4, 7, and 10), confocal laser FITC (2, 5, 8, and 11), or phase-contrast/FITC merged images (3, 6, 9, and 12). Images 1 to 3, untreated(NS) JEG-3 cells; images 4 to 6, GnRH I (100 nmol/L, 24 hours). All cells have been exposed while live to an annexin V recombinant protein conjugated to the FITC fluorophore.In images 2 and 5, no annexin V-reactive PS is present. Images 10 to 12, cells exposed to GnRH I for 48 hours; images 7 to 9 are unstimulated (NS) contemporaneous controls. Noannexin V reactivity is seen in image 8 (unstimulated), whereas GnRH stimulation (48 hours) induces PS translocation and annexin V reactivity. Similar expression of annexinV-FITC–reactive PS was also demonstrated in BPH-1 cells exposed to GnRH I (100 nmol/L) for 48 hours (A, 13–15). GnRH I (100 nmol/L) enhanced expression of proapoptoticproteases cleaved caspase 3 and pro-caspase 3 in both JEG-3 cells (B�D) and BPH-1 cells (E�G) in whole-cell (w.c.) lysates. Representative Western blots of cleaved and pro-caspase-3levels in JEG-3 and BPH-1 cells are shown in B and E, respectively. The histograms in C, D, F, and G represent mean � SE. Mean data were gathered from three independentexperiments.

7541

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 10: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

peripheral reproductive tissues suggested that the receptor present inthe peripheral sites was different from that in the gonadotrope. Recentevidence, however, has suggested that the GnRH receptors present atthese two sites are indeed the same, despite their different signalingbehavior (46–48). A distinction between GnRH signaling in periph-eral compared with pituitary sites also extended to the effects ofGnRH agonist and antagonist analogs. Hence, proliferation of bothendometrial and ovarian cancer cells can be inhibited by both ago-nistic and certain antagonistic analogs of GnRH (10). A solution tothis problem was proposed by Imai et al. (21), who speculated thatG�i coupling of the GnRH receptor to its effectors may be responsiblefor the differences in GnRH agonist/antagonist response betweenperipheral tumors and the anterior pituitary. Our findings support thisconclusion. Interestingly, we have shown that a functional LPA-mediated G�q-coupling activity is extant in these tumor cells; thus, apathophysiological loss of G�q protein cannot explain the paradoxicalchange in GnRH receptor signaling. In the present study, we observednegligible GnRH-mediated activation of G�q in JEG-3 and BPH-1cells. However, other reports have demonstrated that in other repro-ductive tumor lines, such as Ishikawa cells, GnRH can induce cou-pling to G�q (19). In the present study, G�q signaling was clearly notinvolved in the antiproliferative effect of GnRH because LPA-medi-ated activation of G�q failed to inhibit cell growth. Thus, it is probablethat the functional signaling complexes associated with the GnRHreceptors in peripheral tumor sites are able to coerce the receptor intospecific G�i coupling and that such specific complexes are not presentin gonadotropes because L�T2 cell proliferation was not inhibited bycontinuous GnRH exposure (data not shown). Whatever the nature ofthe protein intermediates responsible for this shift in functionality, itis clear that additional receptor-interacting factors can dramaticallyalter the way in which a given ligand can direct its signal to theintracellular environment and eventually induce distinct physiologicend points. This cell environment-specific differential coupling of thereceptor therefore necessitates a reevaluation of terminology withrespect to the nature of ligand interaction with the GnRH receptor atthese peripheral tumor sites. Thus we have shown that whereasantagonist 135-25 can be adequately described as an antagonist at the

anterior pituitary level with respect to GnRH-induced activation of theG�q-based signaling mechanisms, it behaves as an agonist in theperipheral cells because it is almost equally as effective as GnRH instimulating the endogenous G�i-coupled type I GnRH receptors toinhibit cell growth. In addition, we have demonstrated that a separa-tion between these two effects at the periphery and the pituitary can beengineered by alteration of the primary sequence of the GnRH peptideligand. Therefore substitution of the single amino acid in Ile inantagonist 135-18 to 1-MePal in antagonist 125-25 resulted in adramatic elevation of potency at the G�i-coupled peripheral GnRHreceptor but did not change its ability to functionally inhibit the actionof GnRH at the pituitary G�q-coupled receptor.

Both agonist and antagonistic GnRH analogs are now widely usedas therapeutics in gynecology, reproductive medicine, and oncology.The mechanisms of action of the majority of these therapeutics arebased on a continuous treatment regime, causing an anterior pituitaryloss of sensitivity to endogenous GnRH. This causes a reduction ingonadotropin secretion, leading to a diminution of circulating sexsteroids. Classical antagonistic GnRH receptor ligands have an ad-vantage over GnRH agonistic peptides due to the fact that they inhibitthe secretion of gonadotropins and reduce sex steroids immediatelyafter first application, achieving more rapid therapeutic effects thanGnRH agonists (49). The repeated exposure to agonistic agents isrequired to induce a functional desensitization of the anterior pituitarygonadotrope. These agonists initially stimulate the reproductive sys-tem, followed by functional desensitization, which takes days toweeks to occur. For conditions such as prostate cancer, GnRH clas-sical antagonist molecules are therefore preferable to agonists becausethey avoid the so-called “flare” of the disorder that occurs in approx-imately 10% to 20% of patients when agonists are given as singleagents (50). Preexisting antagonistic therapies for reproductive tumorsinvolve the administration of cetrorelix, which has been demonstratedin some circumstances to attenuate the growth of androgen-dependentprostate cancers (51, 52). Interestingly, higher doses of cetrorelix havebeen reported to attenuate the growth of androgen-resistant tumorcells (53, 54), implicating a direct antiproliferative effect, whichrequires a higher dose than that required to inhibit anterior pituitary

Fig. 10. Inhibition of GnRH I-mediated apoptosis by inhibition of JNK or p38 SAPKs. A depicts phase-contrast and confocal microscope images of JEG-3 cells stimulated with100 nmol/L GnRH I (48 hours) in the presence or absence of SB203580 (20 �mol/L). This dose of SB203580 effectively abrogates the ability of GnRH I to stimulate JNK in thesecells but does not significantly affect the activation of ERK1/2 MAPKs (data not shown). Unstimulated cells (1–3) demonstrate no annexin V-FITC staining, whereas with GnRH Itreatment, the anti-annexin V-FITC reactivity of cells (4–6) is apparent. Cotreatment with SB203580 abolishes GnRH-induced generation of annexin V-FITC reactivity (7–9). In B,a similar experimental approach was used for BPH-1 cells. As with A, SB203580 cotreatment with GnRH I abolished the GnRH-induced generation of annexin V-FITC reactivity(compare 4–6 with 7–9).

7542

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 11: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

gonadotropin release, for the classical mode of action of depriving theandrogen-sensitive tumor of steroid. Although deprivation of andro-gen by GnRH analogs is generally beneficial to patients with andro-gen-dependent prostatic cancer, the tumors can often “escape” andreturn as aggressive androgen-independent forms. Thus, there wouldappear to be value in therapies involving GnRH analogs with directantiproliferative effects. In our hands, cetrorelix was significantly lesspotent than antagonist 135-25 when we compared their ability tostimulate PTX-sensitive/G�i-dependent MAPK isoform activation(Fig. 8). In addition, it appears that cetrorelix may not possess aparticularly profound antiproliferative effect on all reproductive tu-mors expressing GnRH receptors, e.g., the antiproliferative effect oftriptorelin (GnRH agonist) on LNCaP prostate cells was actuallyinhibited in the presence of cetrorelix (55). Thus, it is possible thatcetrorelix, which possesses a significantly lower antiproliferative po-tency than triptorelin, acted as a functional antagonist of the agonistaction. In additional experimental paradigms, other classical GnRHantagonists, e.g., antide, have been shown to functionally inhibit theantiproliferative effects of classical GnRH receptor agonists (56).Thus, it is possible that the majority of existing GnRH antagonisttherapeutic agents may not have a significantly potent direct antitumoreffect, which would be desirable for steroid-resistant neoplasms. Theirpoor potency may be due to their poor ability to stabilize/induce theG�i-preferring conformation/state of the type I GnRH receptor. Wehave therefore shown that antagonists can be identified that haveenhanced direct antiproliferative activity via their ability to potentlyactivate the G�i-type of GnRH receptor signaling seen in peripheralreproductive tumors (Figs. 3 and 7). An agent such as antagonist135-25 would theoretically display several properties making it supe-rior to current GnRH-based peptide treatment of reproductive neo-plasms: Firstly, because it is not a pituitary agonist, there is no initialdisease flare (49), whereas its inhibitory action at the pituitary willdecrease serum levels of sex steroids, thereby attenuating steroid-sensitive neoplasm growth. Secondly, its enhanced direct antitumoreffect would be directly cytotoxic to steroid-resistant cells potentiallypresent in the neoplasm. Our elaboration of the principles for G�q

signaling inhibition, diminution of sex steroids, and G�i activation fordirect antiproliferative effects sets the scene for specific developmentof analogs with single or combined effects for the most appropriatetherapy of reproductive tumors.

REFERENCES

1. Schally AV, Arimura A, Baba Y, et al. Isolation and properties of the FSH andLH-releasing hormone. Biochem Biophys Res Commun 1971;43:393–9.

2. Matsuo H, Baba Y, Nair RM, Arimura A, Schally AV. Structure of the porcine LH-and FSH-releasing hormone. I. The proposed amino acid sequence. Biochem BiophysRes Commun 1971;43:1334–9.

3. Baba Y, Matsuo H, Schally AV. Structure of the porcine LH- and FSH-releasinghormone. II. Confirmation of the proposed structure by conventional sequentialanalyses. Biochem Biophys Res Commun 1971;44:459–63.

4. Okon E, Koch Y. Localisation of gonadotropin-releasing hormone in the circumven-tricular organs of human brain. Nature (Lond) 1977;268:445–7.

5. Seeburg PH, Adelman JP. Characterization of cDNA for precursor of human lutein-izing hormone releasing hormone. Nature (Lond) 1984;311:666–8.

6. Oikawa M, Dargan C, Ny T, Hsueh AJ. Expression of gonadotropin-releasinghormone and prothymosin-alpha messenger ribonucleic acid in the ovary. Endocri-nology 1990;127:2350–6.

7. Palmon A, Ben Aroya N, Tel-Or S, et al. The gene for the neuropeptide gonadotropin-releasing hormone is expressed in the mammary gland of lactating rats. Proc NatlAcad Sci USA 1994;91:4994–6.

8. Azad N, Emanuele NV, Halloran MM, Tentler J, Kelley MR. Presence of luteinizinghormone-releasing hormone (LHRH) mRNA in rat spleen lymphocytes. Endocrinol-ogy 1991;128:1679–81.

9. Millar RP, Lu Z-L, Pawson AJ, et al. Gonadotropin-releasing hormone receptors.Endocr Rev 2004;25:235–75.

10. Emons G, Ortmann O, Schulz K, Schally AV. Growth-inhibitory actions of analoguesof luteinizing hormone-releasing hormone on tumor cells. Trends Endocrinol Metab1997;8:355–62.

11. Miller WR, Scott WN, Morris R, Fraser HM, Sharpe RM. Growth of human breastcancer cells inhibited by a luteinizing hormone-releasing hormone agonist. Nature(Lond) 1985;313:231–3.

12. Eidne KA, Flanagan CA, Millar RP. Gonadotropin-releasing hormone binding sites inhuman breast carcinoma. Science (Wash DC) 1985;229:989–91.

13. Huang SC, Tang MJ, Hsu KF, Cheng YM, Chou CY. Fas and its ligand, caspases, andbcl-2 expression in gonadotropin-releasing hormone agonist-treated uterine leiomy-oma. J Clin Endocrinol Metab 2002;87:4580–6.

14. Emons G, Weiss S, Ortmann O, Grundker C, Schulz KD. LHRH might act as anegative autocrine regulator of proliferation of human ovarian cancer. Eur J Endo-crinol 2000;142:665–70.

15. Moretti RM, Marelli MM, Dondi D, et al. Luteinizing hormone-releasing hormoneagonists interfere with the stimulatory actions of epidermal growth factor in humanprostatic cancer cell lines, LNCaP and DU 145. J Clin Endocrinol Metab 1996;81:3930–7.

16. Emons G, Schroder B, Ortmann O, et al. High affinity binding and direct antiprolif-erative effects of luteinizing hormone-releasing hormone analogs in human endome-trial cancer cell lines. J Clin Endocrinol Metab 1993;77:1458–64.

17. Emons G, Ortmann O, Becker M, et al. High affinity binding and direct antiprolif-erative effects of LHRH analogues in human ovarian cancer cell lines. Cancer Res1993;53:5439–46.

18. Schally AV, Comaru-Schally AM, Nagy A, et al. Hypothalamic hormones andcancer. Front Neuroendocrinol 2001;22:248–91.

19. Grundker C, Volker P, Emons G. Antiproliferative signaling of luteinizing hormone-releasing hormone in human endometrial and ovarian cancer cells through G proteinalpha(I)-mediated activation of phosphotyrosine phosphatase. Endocrinology 2001;142:2369–80.

20. Grundker C, Gunthert AR, Westphalen S, Emons G. Biology of the gonadotropin-releasing hormone system in gynecological cancers. Eur J Endocrinol 2002;146:1–14.

21. Imai A, Takagi H, Furui T, et al. Evidence for coupling of phosphotyrosine phos-phatase to gonadotropin-releasing hormone receptor in ovarian carcinoma membrane.Cancer (Phila) 1996;77:132–7.

22. Limonta P, Moretti RM, Marelli MM, et al. The luteinizing hormone-releasinghormone receptor in human prostate cancer cells: messenger ribonucleic acid expres-sion, molecular size, and signal transduction pathway. Endocrinology 1999;140:5250–6.

23. Millar RP, Lowe S, Conklin D, et al. A novel mammalian receptor for the evolu-tionarily conserved type II GnRH. Proc Natl Acad Sci USA 2001;98:9636–41.

24. Neill JD, Duck LW, Sellers JC, Musgrove LC. A gonadotropin-releasing hormone(GnRH) receptor specific for GnRH II in primates. Biochem Biophys Res Commun2001;282:1012–8.

25. Millar RP. GnRH II and type II GnRH receptors. Trends Endocrinol Metab 2003;14:35–43.

26. Pawson AJ, Morgan K, Maudsley SR, Millar RP. Type II Gonadotropin-releasinghormone (GnRH-II) in reproductive biology. Reproduction 2003;126:271–8.

27. Kakar SS, Grizzle WE, Neill JD. The nucleotide sequences of human GnRH receptorsin breast and ovarian tumors are identical with that found in pituitary. Mol CellEndocrinol 1994;106:145–9.

28. Nathwani PS, Kang SK, Cheng KW, Choi KC, Leung PC. Regulation of gonado-tropin-releasing hormone and its receptor gene expression by 17beta-estradiol incultured human granulosa-luteal cells. Endocrinology 2000;141:1754–63.

29. Millar R, Davidson L, Morgan K, et al Mechanism of anti-proliferative and apoptoticeffects of GnRH agonists and antagonists in hormone dependent cancer cell lines.Gynecological Endocrinology 2003; 17 (Supplement 1):3.

30. Kenakin T. Agonist-receptor efficacy. II. Agonist trafficking of receptor signals.Trends Pharmacol Sci 1995;16:232–8.

31. Kenakin T. Ligand-selective receptor conformations revisited: the promise and theproblem. Trends Pharmacol Sci 2003;24:346–54.

32. Brink CB, Wade SM, Neubig RR. Agonist-directed trafficking of porcine alpha(2A)-adrenergic receptor signaling in Chinese hamster ovary cells: L-isoproterenol selec-tively activates G(s). J Pharmacol Exp Ther 2000;294:539–47.

33. Berg KA, Maayani S, Goldfarb J, et al. Effector pathway-dependent relative efficacyat serotonin type 2A and 2C receptors: evidence for agonist-directed trafficking ofreceptor stimulus. Mol Pharmacol 1998;54:94–104.

34. Gerhardt CC, Gros J, Strosberg AD, Issad T. Stimulation of the extracellular signal-regulated kinase 1/2 pathway by human beta-3 adrenergic receptor: new pharmaco-logical profile and mechanism of activation. Mol Pharmacol 1999;55:255–62.

35. Gong H, Sun H, Koch WJ, et al. Specific beta(2)AR blocker ICI 118,551 activelydecreases contraction through a G(i)-coupled form of the beta(2)AR in myocytes fromfailing human heart. Circulation 2002;105:2497–503.

36. Anderson L, McGregor A, Cook JV, Chilvers E, Eidne KA. Rapid desensitization ofGnRH-stimulated intracellular signalling events in alpha T3-1 and HEK-293 cellsexpressing the GnRH receptor. Endocrinology 1995;136:5228–31.

37. Maudsley S, Pierce KL, Zamah AM, et al. The beta(2)-adrenergic receptor mediatesextracellular signal-regulated kinase activation via assembly of a multi-receptorcomplex with the epidermal growth factor receptor. J Biol Chem 2000;275:9572–80.

38. Ott TR, Troskie BE, Roeske RW, et al. Two mutations in extracellular loop 2 of thehuman GnRH receptor convert an antagonist to an agonist. Mol Endocrinol 2002;16:1079–88.

39. Morgan K, Conklin D, Pawson AJ, et al. A transcriptionally active human type IIgonadotropin-releasing hormone receptor gene homolog overlaps two genes in theantisense orientation on chromosome 1q.12. Endocrinology 2003;144:423–36.

7543

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 12: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

40. Kang SK, Tai CJ, Cheng KW, Leung PC. Gonadotropin-releasing hormone activatesmitogen-activated protein kinase in human ovarian and placental cells. Mol CellEndocrinol 2000;170:143–51.

41. Kimura A, Ohmichi M, Kurachi H, et al. Role of mitogen-activated protein kinase/extracellular signal-regulated kinase cascade in gonadotropin-releasing hormone-induced growth inhibition of a human ovarian cancer cell line. Cancer Res 1999;59:5133–42.

42. Martin SJ, Reutelingsperger CP, McGahon AJ, et al. Early redistribution of plasmamembrane phosphatidylserine is a general feature of apoptosis regardless of theinitiating stimulus: inhibition by overexpression of Bcl-2 and Abl. J Exp Med1995;182:1545–56.

43. Beauparlant P, Shore GC. Therapeutic activation of caspases in cancer: a question ofselectivity. Curr Opin Drug Discov Dev 2003;6:179–87.

44. Mangoura D, Pelletiere C, Leung S, Sakellaridis N, Wang DX. Prolactin concurrentlyactivates src-PLD and JAK/Stat signaling pathways to induce proliferation whilepromoting differentiation in embryonic astrocytes. Int J Dev Neurosci 2000;18:693–704.

45. Miles LE, Hanyaloglu AC, Dromey JR, Pfleger KD, Eidne KA. Gonadotropin-releasing hormone receptor-mediated growth suppression of immortalized LbetaT2gonadotrope and stable HEK293 cell lines. Endocrinology 2004;145:194–204.

46. Chegini N, Rong H, Dou Q, Kipersztok S, Williams RS. Gonadotropin-releasinghormone (GnRH) and GnRH receptor gene expression in human myometrium andleiomyomata and the direct action of GnRH analogs on myometrial smooth musclecells and interaction with ovarian steroids in vitro. J Clin Endocrinol Metab 1996;81:3215–21.

47. Yin H, Cheng KW, Hwa HL, et al. Expression of the messenger RNA for gonado-tropin-releasing hormone and its receptor in human cancer cell lines. Life Sci1998;62:2015–23.

48. Chatzaki E, Bax CM, Eidne KA, et al. The expression of gonadotropin-releasinghormone and its receptor in endometrial cancer, and its relevance as an autocrinegrowth factor. Cancer Res 1996;56:2059–65.

49. Schally AV. Luteinizing hormone-releasing hormone analogs: their impact on thecontrol of tumorigenesis. Peptides 1999;20:1247–62.

50. Schally AV, Comaru-Schally AM. Hypothalamic and other peptide hormones. In:Holland JF, Frei E III, Bast RC Jr, et al. (eds). Cancer medicine, 4th ed. Baltimore,MD: Williams & Wilkins; 1997. p. 1067–86.

51. Redding TW, Schally AV, Radulovic S, et al. Sustained release formulations ofluteinizing hormone-releasing hormone antagonist SB-75 inhibit proliferation andenhance apoptotic cell death of human prostate carcinoma (PC-82) in male nude mice.Cancer Res 1992;52:2538–44.

52. Korkut E, Comaru-Schally AM, Schally AV, Plowman PN. LHRH analogues andbreast cancer. Br J Cancer 1991;64:1190.

53. Jungwirth A, Galvan G, Pinski J, et al. Luteinizing hormone-releasing hormoneantagonist cetrorelix (SB-75) and bombesin antagonist RC-3940-II inhibit the growth ofandrogen-independent PC-3 prostate cancer in nude mice. Prostate 1997;32:164–72.

54. Jungwirth A, Pinski J, Galvan G, et al. Inhibition of growth of androgen-independentDU-145 prostate cancer in vivo by luteinising hormone-releasing hormone antagonistcetrorelix and bombesin antagonists RC-3940-II and RC-3950-II. Eur J Cancer1997;33:1141–8.

55. Ravenna L, Salvatori L, Morrone S, et al. Effects of triptorelin, a gonadotropin-releasing hormone agonist, on the human prostatic cell lines PC3 and LNCaP. JAndrol 2000;21:549–57.

56. Kang SK, Choi KC, Cheng KW, et al. Role of gonadotropin-releasing hormone as anautocrine growth factor in human ovarian surface epithelium. Endocrinology 2000;141:72–80.

7544

GnRH RECEPTOR-MEDIATED INHIBITION OF PROLIFERATION

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

Page 13: Gonadotropin-Releasing Hormone (GnRH) Antagonists Promote ... · Gonadotropin-releasing hormone (GnRH) receptor agonists are exten-sively used in the treatment of sex hormone-dependent

2004;64:7533-7544. Cancer Res   Stuart Maudsley, Lindsay Davidson, Adam J. Pawson, et al.   GnRH Receptor

-Coupling State of the Type IiαTumor Cells by Activating a GPromote Proapoptotic Signaling in Peripheral Reproductive Gonadotropin-Releasing Hormone (GnRH) Antagonists

  Updated version

  http://cancerres.aacrjournals.org/content/64/20/7533

Access the most recent version of this article at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/64/20/7533.full#ref-list-1

This article cites 53 articles, 16 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/64/20/7533.full#related-urls

This article has been cited by 12 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  SubscriptionsReprints and

  [email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://cancerres.aacrjournals.org/content/64/20/7533To request permission to re-use all or part of this article, use this link

Research. on October 1, 2020. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from


Recommended