Presence of gonadotropin-releasing hormone and its messenger ribonucleic acid in human ovarian epithelial carcinoma

Presence of gonadotropin-releasing hormone and its messenger ribonucleic acid in human ovarian epithelial carcinoma Tsukasa Ohno, MD, Atsushi Imai, MD, PhD, Tatsuro Furui, MD, Kyoko Takahashi, MD, and Teruhiko Tamaya, MD, PhD

GijU,japan OBJECTIVE: The purpose of this study was to investigate the expression of gonadotropin-releasing hormone messenger ribonucleic acid and the presence of gonadotropin-releasing hormone in human ovarian carcinoma known to have gonadotropin-releasing hormone binding sites and to be affected by gonadotropin-releasing hormone analog. STUDY DESIGN: Human ovarian carcinomas surgically removed and human ovarian carcinoma cell lines were examined. Gonadotropin-releasing hormone was determined by a radioimmunoassay and a bioassy. Gonadotropin-releasing hormone messenger ribonucleic acid was determined by reverse transcription polymerase chain reaction using oligonucleotide primers synthesized according to the published human gonadotropin-releasing hormone sequence. RESULTS: Gonadotropin-releasing hormone was shown to be present in extracts of ovarian mucinous cystadenocarcinoma sample (0.8 ± 0.12 pg/mg of protein) and ovarian adenocarcinoma cell line SK-OV3 (0.92 ± 0.17 pg/mg of protein) but not in the normal ovary and placenta. Two of two extract samples from individual cases evoked dose-dependent phosphoinositide breakdown in rat granulosa cells similar to that caused by authentic gonadotropin-releasing hormone. Gonadotropin-releasing hormone messenger ribonucleic acid was detected in two of two mucinous cystadenocarcinoma specimens, one of one serous cystadenocarcinoma, and SK-OV3 cells but not in the dysgerminoma, mucinous cystadenoma, and normal ovary and placenta. CONCLUSION: The demonstration of gonadotropin-releasing hormone and its messenger ribonucleic acid raises the possibility that gonadotropin-releasing hormone may play an autocrine regulatory role in the growth of ovarian carcinoma. (AM J OBSTET GVNECOL 1993; 169:605-1 0.)

Key words: Human ovarian carcinoma, gonadotropin-releasing hormone, gonadotropin-releasing hormone messenger ribonucleic acid In addition to its classic hypophysiotropic action, gonadotropin-releasing hormone (CnRB) might playa role as a modulator of activity in the brain and many peripheral organs, including the ovary.I.3 GnRH analogs have been used in the therapy of certain hormonesensitive cancers such as breast cancer:' 5 prostate cancer,6. 7 and ovarian cancer. 8 • 9 Growth of these tumors is inhibited by GnRH analog, and clinical studies have indicated beneficial effects even in the treatment of some metastatic tumors. The antitumor action of GnRH analog is presumed to result from a desensitization or From the Department of Obstetrics and Gynecology, Gifu University School of Medicine. Supported by research grants Nos. 01570923 and 04670996 from the Ministry of Education, Culture, and Science, Japan. Received for publication January 11, 1993; revised April 5, 1993; accepted April 13, 1993. Reprint requests: Atsushi Imai, MD, PhD, Department of Obstetrics and Gynecology, Gifu University School of Medicine, Tsukasamachi, Gifu 500 Japan. Copyright © 1993 by Mosby-Year Book, Inc.

0002-9378193 $1.00 + .20 6!1147915

down-regulation of GnRH receptors in the pituitary, with a consequent decline in gondaotropin secretion and gonadal hormone production. There are, however, indications that GnRH analog can also suppress the growth of the tumor cells in vitro lO - 12 and that the presence of specific binding sites for GnRH is demonstrated in the tumors responsive to CnRH analog. II -18 These findings suggest direct regulatory effects of GnRH on tumor growth. Because of the short half-life, GnRH is undetectable or very low in the general circulation. It is unlikely that GnRH of hypothalamic origin could reach the peripheral organs in sufficient concentrations to activate its receptors and exert a direct action, although GnRH is supposedly secreted in a pulsatile fashion from the hypothalamus. Consideration of these findings lead to the possibility that GnRB might be produced locally within these tumors. We therefore examined human ovarian carcinomas and ovarian carcinoma cell lines for the presence of GnRH and its messenger ribonucleic acid (mRNA).

605

606

Ohno et al.

Material and methods

Materials. myo[2- 3 H]Inositol (740 GBq!mmol) and [3,4-'H]GnRH (2.22 TBq!mmol) were obtained from New England Nuclear (Boston). Anti-GnRH and antirabbit gamma globulin antibodies were purchased from Chemic on and Medac. Authentic GnRH was from Sigma (St. Louis) and buserelin acetate was a gift from HoechstJapan, respectively. Moloney murine leukemia reverse transcriptase and Taq polymerase were from Perkin-Elmer-Cetus. Restriction endonuclease enzymes were products of Takara Shuzou. The mRNA purification kit was a product of Pharmacia-LKB. All other chemicals were of experimental reagent grade. Cells and cell cultures. Human breast carcinoma cell line ZR7S-l, human endometrial carcinoma cell line RL9S-2, human ovarian adenocarcinoma cell line SKOV3, human placental cell line 3A (tPA-30-1), human endometrioid carcinoma cell line MS7Sl, and human uterine cervical carcinoma cell line SiHa were obtained from American Type Culture collection. Human endometrial carcinoma cell line HHUA was from Dr. Kuramoto, Kitasato University Qapan). The cells were grown at 37° C in an appropriate medium containing 10% fetal bovine serum in a ISO em" flask in a humidified atmosphere of S% carbon dioxide and 9S% air. Tissue collection. Ovarian tumors and normal tissues were placed in ice-cold phosphate-buffered saline solution immediately after surgical removal, and representative portions were excised to prepare the material for histologic frozen sections. These tissue samples were washed and immediately used or stored in liquid nitrogen. Peptide extraction. Tissues or cell pellets were extracted according to standard techniques.)9 In brief, a volume of acidic ethanol (SO% ethanol-acetic acid, 2 mol!L) was added, equal to four to five times the volume of the tissues or cell pellets. They were subsequently homogenized at 0° C with a Teflon homogenizer. The extracts were clarified by centrifugation at 100,000g for 1 hour. The supernatant was lyophilized and stored at - 80° C. Protein content was determined by the method of Lowry et al. 20 with bovine serum albumin as a standard. GnRH radioimmunoassay (RIA). GnRH peptide in the acidic ethanolic extract was measured with antiserum in an established RIA.2) The specimens were reconstituted in phosphate-buffered saline solution containing ethylenediamine tetraacetic acid, 1 mmol!L, and 0.1 % bovine serum albumin. The method included addition to all tubes of SOO fLl of various concentrations of extracts (or standards) diluted with the phosphate buffered saline-ethylenediaminetetraacetic acid-bovine serum albumin solution, SOO fLl of anti-GnRH serum at dilutions of 1 : SOOO to 1: SO,OOO, and S fLl of tritiated GnRH (2 x 106 disintegrations/min, 16.6 pmol). All tubes were incubated for 12 hours at 4°C, followed by

September 1993 Am J Obstet Gynecol

addition of 200 fLl of approximately diluted antirabbit gamma globulin. Additional incubations were carried out at 4° C for 16 hours, after which the immune complexes were trapped by membrane filter, washed, and counted. Standard curves ranging from 0.1 pg to 1 ng were included in each assay, using authentic GnRH or analog buserelin. The lower limits of sensitivity of these assays for SOO fLl samples ranged from 0.03 pg to O.OS pg of GnRH. These values were assigned to the authentic GnRH and buserelin. In some experiments, to validate the use of the anti-GnRH antibody from Chemicon to determine GnRH concentration in our RIA, comparisons of assay value were made with the different antisera from Chemicon and Medac. The assay value was found not to differ between the two antisera with the same extract preparation. This indicated that the assay could be applied to titrate a immunoreactive GnRH In our samples. GnRH-like activity assay. GnRH and its analog stimulate phosphoinositide metabolism in rat granulosa cells. 22 . 23 We applied this system to examine GnRH-like activity in the peptide extracted from ovarian carcinomas and normal tissues. Granulosa cells were prepared from the ovaries of immature rats, as described previously.22. 23 Phosphoinositide metabolism was observed by formation if inositol phosphates from tritiated inositol-labeled granulosa cells. The cells were prelabeled with tritiated inositol (37 MBq!ml of medium) for 2 days, a time sufficient to reach isotopically steady state. The cells were washed and resuspended in balanced salt solution (sodium chloride 135 mmol!L, potassium chloride 4.S mmol!L, magnesium chloride O.S mmol!L, calcium chloride I.S mmol!L, glucose S.6 mmol!L, N -[2 -hydroxyethy I]piperazine-N -[2 -ethane sulfonic acid] 10 mmol!L, pH 7.4) (2 x 106 cells/ml) containing lithium chloride 10 mmol!L and then exposed to the various concentrations of the extract or authentic GnRH. The incubations were terminated by adding four volumes of chloroform methanol-hydrochloric acid (1 : 2 : 0.001, vol!vol!vol), followed by one volume of chloroform and one of water. Tritiated inositol sugars were analyzed by anion exchange column chromatography.22.23 I

Granulosa cell viability was estimated by determining the percentage of trypan blue-excluding cells. Treatment with the extracts tested did not preferentially affect viability, compared with nonexposed cells. RNA isolation. Total RNA was extracted from tissues or cell pellets according to the procedure of Chromcynski and Sacchi.24 Briefly, the specimens were homogenized in guanidium thiocyanate followed by acid phenol-chloroform extraction, precipitated with isopropanol, and reextracted with guanidium thiocyanate and isopropanol precipitation. Poly(A) mRNA was then pre-

Ohno et al.

Volume 169, Number 3 Am J Obstet Gynecol

pared with oligo(dT) cellulose columns/' according to the manufacturer's instructions. Reverse transcription and polymerase chain reaction amplification. A random primed complementary DNA (cDNA) library was obtained from 200 ng of each of the poly(A) RNA prepared with Moloney murine leukemia reverse transcriptase under the conditions recommended by the supplier: The cDNA reaction (25 ]J.l) was diluted with 300 ]J.l of water, heat denatured at 95° C for 5 minutes, and quickly chilled on ice. The eDNA (I ]J.I) was amplified in a 50 ]J.I reaction containing Tris-hydrochloric acid 10 mmolJL (pH 8.3), potassium chloride 50 mmolJL magnesium chloride 2 mmolJL, 50 ]J.mollL each deoxynucleotide triphosphates, 1.25 U of Tag polymerase, and primer 0.05 ]J.mollL. 26 The sequences of oligonucleotide primers, synthesized according to the published human GnRH sequence!7 were as follows: primer a 5'-TGGAAGGCTGCTCCAGCCAG-3', primer b 5' -TCCTTCTGGCCCAATGGATT-3', primer c 5'-ACAACACAGCACTTTATTAT-3'. The predicted fragments amplified by polyermase chain reaction were 245 and 380 bp; one tube contained primers a and b with a predicted DNA fragment of 245 bp, and the other tube contained primers a and c with a predicted fragment of 380 bp, We carried out 35 cycles of amplification: denaturation at 94° C for 2 minutes, annealing at 55° C for 1 minute, and extension 72° C for 2 minutes, followed by a final extension for 10 minutes at 72° C. The DNA product (IO ]J.l) was run on 2% agarose gels, and bands were visualized by ethidium bromide staining on an ultraviolet transilluminator. Restriction enzyme analysis. Polymerase chain reaction products were digested with restriction endonuclease enzymes TthHB8I andAfi II under the conditions recommended by the supplier. The digested products, along with the untreated aliquots of each polymerase chain reaction sample, were then fractionated by electrophoresis in 2% agarose gels and stained. Statistics. Statistical analysis was performed by t test. Differences were considered significant if p < 0.01. Results

GnRH immunoreactive peptides. Fig. 1 shows cross reaction of extracts from human ovarian carcinoma with antiserum to GnRH. Acetic acid-ethanol extract of the ovarian mucinous cystadenocarcinoma caused a highly significant and dose-dependent inhibition of specific tritiated GnRH binding to antiserum. The slope of the inhibition curve produced by serial dilutions of the extract was similar to those of authentic GnRH or analog buserelin. In addition, equivalent amounts of extracts from human ovarian carcinoma cell line, SKOV3, reacted with antiserum. GnRH contents in mucinous cystadenocarcinoma removed from a certain case and SK-OV3 cells ranged from 0.8 ± 0.12 and 0.92 ±

Equivalent of Tissues ( 0.1

0.3

1

607

) (g protein) 3

10

i

:r:

q:

~ '"TI :r:

* ..,

'0.

0.5

.~

c.

o

<:

:::J

E E

o

0.1

1.0

10

Gn-RH or Buserelin (------

100 (pmol)

Fig. 1. Effects of extracts of ovarian carcinoma, normal ovary, and placenta on binding of tritiated GnRH by a GnRH-specific antiserum. Antiserum dilution of 1 : 50,000 was incubated with tritiated GnRH (2 x 10 6 disintegrations/min) in the presence of peptide fractions extracted from increasing amount of mucinous cystadenocarcinoma (0), ovarian adenocarcinoma cell line SK-OV3 (-), normal ovary (t:.), and placenta ( .. ) for 12 hours at 0° C. Standard curves ranging from 0.1 pg to I ng were included in each assay, using authentic GnRH (0) or analog buserelin acetate (e). Results are means ± SD of three determinations.

0.17 pg/mg protein, respectively. No immunoreactivity was detectable in extracts of the human ovary and term placenta. GnRH-like activity. The biologic activity of the GnRH-like peptide was evaluated in a rat granulosa cell assay previously shown to be sensitive to GnRH.22. 23 The granulosa cells respond to GnRH to produce inositol phosphates as a result of phosphoinositide breakdown. On exposure to the acidic ethanol extracts obtained from ovarian mucinous cystadenocarcinoma of individual two cases, both cases induced remarkable production of inositol phosphate production in a dosedependent manner, as shown in Fig. 2. GnRH (100 nmollL) also stimulated phosphoinositide breakdown. The extracts from ovary or placenta, however, had no effects; inositol phosphate production rates caused by the extracts obtained from 10 gm of ovary and 8 gm of placenta were 105% ± 12% and 93% ± 17% of control, respectively. GnRH messenger ribonucl~ic acid. Polymerase chain reaction amplification of first-strand eDNA from human ovarian tumors, including two samples of mucinous cystadenocarcinoma, two of serous cystadenocarcinoma, one of dysgerminoma, one of mucinous cystadenoma, and the normal ovary and placenta was conducted with two sets of oligonucleotide primers, as

608

Ohno et al.

September 1993 Am J Obstet Gynecol

(b)

(a)

gc o

HH UA RL95 2

o 200 '0

...,

I\)

"-

..'"

SK OV-3 ZR75-1

(11

3A (t PA-30-J)

0)

MS751

....

SIHA HHUA RL95-2 SK OV-3 ZR 75-1

CD Q)

c 0

z

I

q: c

(')

'cE" 0

'Ec" 0

-Q

"Q

0 c

~

'" '"

-~

> 0

a

0.001 0.01

0.1

1.0

Equivalent of Tissue (g protein)

(J)

0

c

'" >

.~

0

Fig. 2. Effects of extracts of ovarian carcinoma on inositol phosphate production in rat granulosa cells. Tritiated inositollabeled rat granulosa cells (1 x 106 cells) were incubated for 10 minutes at 37° C with authentic GnRH 100 nmoVL (a), the peptide fractions extracted from 1.0 gm of ovarian mucinous cystadenocarcinoma of two cases (b), and peptide fraction extracted from increasing amounts of ovarian mucinous cystadenocarcinoma (e). Each point represents mean ± SD of triplicate determinations. Asterisk, p < 0.005 versus none.

I\)

3A (tPA-30-J)

'"

MS751 SIHA

A

NW ","00

(J'IO

Fig. 4. Polymerase chain reaction amplification of first-strand cDNA from various human tumor cell lines. Oligonucleotide primers a and b were used in lanes 1 through 7, and primers a and c were used in lanes 8 through 14. Gels were stained with ethidium bromide, and bands were visualized with ultraviolet light. Templates poly(A) RNA from HHUA cells (lanes 1 and 8), PL95-2 cells (lanes 2 and 9), SK-OV3 cells (lanes 3 and 10), ZR 75-1 cells (lanes 4 and 11), 3A(tPA-30-1) cells (lanes 5 and 12), MS 751 cells (lanes 6 and 13), and SiHa cells (lanes 7 and 14).

658 489 267 -

1234567891011121314

Fig. 3. Polymerase chain reaction amplification of first-strand cDNA from various human ovarian carcinoma and normal tissues. Oligonucleotide primers a and b were used in lanes 1 through 7, and primers a and c were used in lanes 8 through 14. Gels were stained with ethidium bromide, and bands were visualized with ultraviolet light. Templates poly(A) RNA from mucinous cystadenocarcinoma (lanes 1 and 8), mucinous cystadenocarcinoma ((lanes 2 and 9), dysgerminoma (lanes 3 and 10), serous cystadenocarcinoma (lanes 4 and 11), mucinous cystadenoma (lanes 5 and 12), normal ovary (lanes 6 and 13), and placenta (lanes 7 and 14).

described in Material methods. As shown in Fig. 3, the ovarian carcinomas tested all yielded expected 245 and 380 bp products, whereas there was no product from dysgerminoma, mucinous cystadenoma, and the normal tissues, ovary, and placenta.

Human breast carcmoma cell line ZR75-1, human endometrial carcinoma cell lines RL95-2 and HHUA, and human ovarian adenocarcinoma cell line SK-OV3 gave a predominant product identical to that obtained in tumor tissues, as shown in Fig. 4. No detectable products were obtained in human placental cell line 3A (tPA-30-1), human endometrioid carcinoma cell line MS751, and human uterine cervical carcinoma cell line SiHa. The predicted sequences of the DNA products generated by the set of GnRH primers contained TthHB81 and Aft II restriction sites. Cleavage of the 380 bp product was expected to yield two fragments of 130 and 250 bp by TthHB8I, two of 113 and 267 bp by Aft II, and three of 113, 130, and 137 bp by combination of the enzymes. Fig. 5 shows that digestion of the polymerase chain reaction products produced fragments of the expected sizes. Comment

These results demonstrate for the first time the mRNA expression of GnRH and the presence of GnRH

Ohno et al.

Volume 169. Number 3 Am J Obstet Gynecol

in human ovarian carcinoma reported to have GnRH binding sites. We detected a GnRH with immunologic properties indistinguishable from those of authentic GnRH in the acetic acid-ethanol extracts from the human ovarian carcinoma tissue and human ovarian carcinoma cell lines (SK-OV3). The possibility that our results could be accounted for by nonspecific immunoreactivity seems highly unlikely because the normal ovary and placenta analyzed in parallel to the ovarian carcinomas were totally negative. We also demonstrated that GnRH in the ovarian carcinoma was biologically active. GnRH and analog are known to bind to GnRH receptor and stimulate phosphoinositide metabolism in granulosa cells 22 . 23 and to suppress gonadotropin-induced maturation and steroidogenesis!8. 29 The peptide extracted from ovarian carcinomas induced the dose-dependent stimulation of inositol phosphate production in rat granulosa cells. The findings of immunoreactivity and bioactivity clearly indicate the presence of GnRH in the ovarian carcinoma, although the concentration (picogram range per milligram of protein) was much lower than those (nanogram range) in the hypothalamus. 3o Others have previously demonstrated the presence of GnRH-like peptide in the normal ovary'" 32 and placenta. " Our failure to detect immunoreactive GnRH in these tissues might be because of low sensitivity of our assay condition; our tritiated RIA was fivefold less sensitive than the iodine 125 RIA for analysis of GnRH. The authenticity of the GnRH and its production by the ovarian carcinoma was further established by the demonstration of GnRH mRNA by polymerase chain reaction amplification study in ovarian carcinoma tissues and ovarian carcinoma cell lines. This demonstration was based on the generation of identical polymerase chain reaction-amplified products to GnRH eDNA from the hypothalamus, a well-established site of GnRH synthesis. The amplification products had the predicted sizes and included complementary sequences for GnRH probes, and restriction enzymes fragments had the predicted sizes. The sizes of the amplification products ruled out that they could have resulted from amplification of genomic DNA. The possibility that our results might be accounted for by cross-tissue contamination seemed highly unlikely, because no reverse transcription polymerase chain reaction products were detected when the GnRH primers were used with ovarian nonepithelial tumor dysgerminoma, benign tumor; normal ovary and other cell lines, and because reverse transcription and polymerase chain reaction blanks always remained negative. Given these conditions of reverse transcription polymerase chain reaction amplification and direct sequential analysis (data not shown) of the polymerase chain reaction-amplified products, we believe that amplified sequences of the eDNA were generated from human ovarian carcinoma GnRH mRNA.

609

380265 260

H~234 Fig. 5. Restriction enzyme analysis of the polymerase chain reaction products from ovarian mucinous cystadenocarcinoma sample. After digestion by TthHB8I, Aft II or TthHB81 plus Aft II, digested and untreated products were size-fractionated on 2% agarose and analyzed by ethidium bromide staining. Untreated (lane 1), TthHB 8I-treated (lane 2), Aft II-treated (lane 3), TthHB81 plus Aft II-treated (lane 4).

We failed to detect GnRH mRNA in the human placenta and ovary, the tissues that have the capacity to produce GnRH. 31 · 33 The most plausbile explanation is that the placental or ovarian GnRB gene might differ from that in hypothalamus. In fact, previous studies isolated and characterized the human GnRH gene in placental tissue, showing a slight difference from the hypothalamic GnRH cDNA. 34 . '5 Emons et a1. 13 and we"" previously reported the existence of GnRB receptors in human ovarian epithelial carcinoma tissue and cell lines. Although the functional role of the receptor in ovarian carcinoma is still obscure, GnRH can stimulate phosphoinositide metabolism through phospholipase C in ovarian carcinoma membrane,36 in such a way as to permit regulation of tumor growth by GnRH. It is unlikely that endogenous hypothalamic GnRH ever reaches concentrations in peripheral blood that are necessary to stimulate ovarian tumor GnRH receptor. The current finding that GnRH might be produced locally within the ovarian carcinoma suggests that GnRH receptor could exert an autocrine role on tumor growth. The demonstration of GnRH production by GnRH-responsive tumor is not without precedent because GnRH has been reported to be expressed in breast carcinoma cells known to have GnRH receptors and to be affected by GnRH analog. 19 In the breast carcinoma it is also suggested that GnRH may serve an autocrine regulatory role. Last, the current data show that human ovarian epithelial carcinoma can produce GnRH known to have GnRH-binding sites. GnRH might act as an autocrine regulator of ovarian carcinoma proliferation, and the relatively high dose of GnRH analog might induce desensitization to GnRH or down-regulation of GnRH receptor with a consequent decline of tumor growth, in

610

Ohno et al.

an analogous manner to GnRH analog's action on anterior pituitary. Thereby it should be interesting to look for a correlation between GnRH response, the histologic condition of the tumors, and the clinical course of the disease. This might be part of a possible point of attack for therapeutic approaches with GnRH analog in this malignancy. REFERENCES 1. Leung PCK, Steele GL. Intracellular signaling in the gonads. Endocrinol Rev 1992;13:476-98. 2. Imai A, Furui T, Tamaya T. Is extrapituitary action of gonadotropin-releasing hormone biologically significant? Ann Clin Biochem 1992;29:477-80. 3. Schally A, Comaru-Schally AM, Redding lW. Antitumor effects of analogs of hypothalamic hormones in endocrinedependent cancers. Proc Soc Exp Bioi Med 1984;175:25981. 4. Harris AL, Carmichael ], Cantwell BML, Dowsett M. Zolandex: endocrine and therapeutic effects in postmenopausal breast cancer. Br] Cancer 1989;59:97-9. 5. Manni A, Santen R, Harvey H, Lipton A, Max D. Treatment of breast cancer with gonadotropin-releasing hormone. Endocrinol Rev 1986;7:89-113. 6. Crawford ED, Eisenberger MA, McLeod DG, et al. A controlled trial of leuprolide with and without flutamide in prostatic carcinoma. N Engl] Med 1989;321:419-24. 7. Schroder FH. Hormonal manipulation of prostatic cancer; too soon for total androgen blockade? BM] 1991;303: 1489-90. 8. Parmar H, Rustin G, Lightman SL, Phillips RH, Hanham IW, Schally AV. Response to o-Trp-6-luteinizing hormone releasing hormone (Decapeptyl) microcapsules in advanced ovarian cancer. BM] 1988;296:1229. 9. Emons G, Ortmann 0, Pahwa GP, Hackenberg R, Oberheuser F, Schulz KD. Intracellular actions of gonadotropic and peptide hormones and the therapeutic value of GnRH agonists in ovarian cancer. Acta Obstet Gynecol Scand 1992;71(suppI155):31-8. 10. Sharoni Y, Bosin E, Miinster A, Levy J, Schally AV. Inhibition of growth of human mammary tumor cells by potent antagonists of luteinizing hormone-releasing hormone. Proc Nat! Acad Sci USA 1989;86:1648-51. 11. Fekete M, Zalatnai A, Comaru-Schally AM, Schally AV. Membrane receptors for pep tides in experimental and human pancreatic cancers. Pancreas 1989;4:521-8. 12. Eidne KA, Flanagan CA, Harris NS, Millar RP. Gonadotropin-releasing hormone (GnRH)-binding sites in human breast cancer cell lines and inhibitory effects of GnRH antagonists.] Clin Endocrinol Metab 1987;64:42532. 13. Emons G, Pahma GS, Brack C, Sturm R, Oberheuser F, Knuppen R. Gonadotropin-releasing hormone binding sites in human epithelial ovarian carcinoma. Eur] Cancer Clin Oncol 1989;25:215-21. 14. Fekete M, Bajusz S, Groot K, Csernus V], Schally AV. Comparison of different agonists and antagonists of luteinizing hormone-releasing hormone for receptorbinding ability to rat pituitary and human breast cancer membrane. 15. Kadar T, Ben-Dayid M, Pontes ]E, Fekete M, Schally AV. Prolactin and luteinizing hormone-releasing hormone receptors in human benign prostatic hyperplasia and prostate cancer. Prostate 1988;12:299-307. 16. Szende B, Srkalovic G, Timar ], et al. Localization of receptors for luteinizing hormone-releasing hormone in pancreatic and mammary cancer cells. Proc Nat! Acad Sci USA 1991;88:4153-6. 17. Srkalovic G, Szende B, Redding lW, Groot K, Schally AV. Receptors for o-Trp6-luteinizing hormone-releasing hor-

September 1993 Am J Obstet Gynecol

18.

19.

20. 21.

22.

23.

24. 25. 26. 27.

28.

29.

30.

31. 32. 33. 34.

35.

36.

mone, somatostatin, and insulin-like growth factor I in MXT mouse mammary carcinoma. Proc Soc Exp Bioi Med 1989; 192:209-18. Srkalovic G, Witt!iff ]L, Schally AV. Detection and partial characterization of receptors for [o-trp6]-luteinizing hormone-releasing hormone and epidermal growth factor in human endometrial carcinoma. Cancer Res 1990;50; 1841-6. Harris N, Dut!ow C, Eidne K, Dong KW, Roberts], Millar R. Gonadotropin-releasing hormone gene expression in MDA-MB-231 and ZR-75-1 breast carcinoma cell line. Cancer Res 1991;51:2577-81. Lowry OH, Rosenbrough N], Farr AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Bioi Chern 1951; 193:265-75. Nett TM, Akbar AM, Niswender GD, Hedlund MT, White WF. A radioimmunoassay for gonadotropin-releasing hormone (GnRH) in serum. ] Clin Endocrinol Metab 1973; 36:880-5. Iida K, Imai A, Tamaya T. Stimulatory effects of estrogen on gonadotropin-releasing hormone-induced phosphoinositide turnover in granulosa cells. ] Steroid Biochem Mol Bioi 1991;38:583-6. Imai A, Iida K, Tamaya T. Tight coupling of gonadotropin-releasing hormone receptor to stimulated phosphoinositide turnover and anti-gonadotropic action. Obstet Gynecol Invest 1992;33:36-41. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-chloroform extraction. Anal Biochem 1987;162:156-9. Aviv H, Leder P. Purification of biologically active globin messenger RNA by chromatography on oligothymidylic acid-cellulose. Proc Nat! Acad Sci USA 1972;69;1408-12. Witsell AL, Schook LB. Clonal analysis of gene expression by PCR. Biotechniques 1990;9:318-22. Adelman]P, Mason A], Hayflick]S, Seeburg PH. Isolation of the gene and the hypothalamic cDNA for the common precursor of gonadotropin-releasing hormone and prolactin release-inhibiting factor in human and rat. Proc Natl Acad Sci USA 1986;83: 179-83. Imai A, Iida K, Tamaya T. Gonadotropin-releasing hormone has a biphasic action on aromatase activity through protein kinase C in granulosa cells. Int J Ferti! 1993;38: 50-6. Wang J, Lee V, Leung PCK. Differential role of protein kinase C in the action of luteinizing hormone-releasing hormone on hormone production in rat ovarian cells. AM ] OBSTET GYNECOL 1989;160:984-9. Wheaton ]E. Regional brain content of luteinizing hormone-releasing hormone in sheep during the estrous cycle, seasonal anestrus, and after ovariectomy. Endocrinology 1979; 104:839-44. Aten RF, Williams AT, Behrman HR. Ovarian gonadotropin-releasing hormone-like protein(s): demonstration and characterization. Endocrinology 1986; 118:961-8. Aten RF, Ireland]], Weems CW, Behrman HR. Presence of gonadotropin-releasing hormone-like proteins in bovine and ovine ovaries. Endocrinology 1987;120:1727-33. Kohdr GS, Siler-Khodr TM. Placental LRF and its synthesis. Science 1980;207:315-7. Petraglia F, Woodruff TK, Botticelli G, et al. Gonadotropin-releasing hormone, inhibin, and activin in human placenta: evidence for a common cellular localization. ] Clin Endocrinol Metab 1992;74:1184-8. Radovick S, Wondisford FE, Nakayama Y, Yamada M, Cutler GB, Weintraub BD. Isolation and characterization of the human gonadotropin-releasing hormone gene in the hypothalamus and placenta. Mol Endocrinol 1990;4: 476-80. Imai A, Ohno T, Furui T, Takahashi K, Matsuda T, Tamaya T. Gonadotropin-releasing hormone stimulates phospholipase C but not protein phosphorylation/dephosphorylation in plasma membrane from human epithelial ovarian cancer. Int] Gynecol Cancer [In press].