Human Reproduction vol.6 no.8 pp. 1063 -1069, 1991

Comparative stimulatory effect of gonadotrophin releasing hormone (GnRH) and GnRH agonist upon pulsatUe human chorionic gonadotrophin secretion in supervised placental explants: reversible inhibition by a GnRH antagonist E.R.Barnea1, M.Kaplan2 and Z.Naor3

Introduction

The University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School at Camden, Cooper Hospital/University Medical Center, Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and Infertility, Camden, NJ, USA, 2Feto-Placental Endocrinology Unit, Rappaport Institute, Technion, P.O. Box 9697, Efron Street, Haifa and 3Department of Biochemistry, Faculty of Life Science, University of Tel Aviv, Ramat Aviv, Israel

Gonadotrophin releasing hormone (GnRH) is a decapeptide which was purified from the hypothalamus in the early seventies (Nekda etal., 1982). Since its identification, its role as the primary regulator of luteinizing hormone (LH) and follicle stimulating hormone (FSH) secretion by the pituitary has been established. Later studies revealed that GnRH secretion is pulsatile, reaching the pituitary via the portal circulation (Pohl et al., 1984), where it acts through specific receptors. Receptors for GnRH have also been described in several extrapituitary organs such as ovary, testis and placenta (Popkin et al., 1983; Currie et al., 1981). The presence and production of this decapeptide was also documented outside the hypothalamus in the placenta (Siler-Khodr and Khodr, 1979), where it was thought to exert a paracrine/autocrine action through specific receptors located on the syncytial cell membrane (Belisle etal., 1984). High doses of GnRH added to placental cells and explants were reported to stimulate the secretion of human chorionic gonadotrophin (HCG) its free alpha subunit, progesterone, oestradiol, oestrone and oestriol (Siler-Khodr et al., 1986). Human chorionic gonadotrophin is a prime marker of placental presence and function in early pregnancy. Episodic HCG secretion in peripheral plasma from pregnant women during early pregnancy has been documented (Owens etal., 1981; Padmanbhan et al., 1989). Two types of pulsatility were noted: long term, occurring every few hours and short term, lasting less than an hour. Since circulating HCG levels undergo major changes during the first trimester (a rapid rise followed by a

' To whom correspondence should be addressed

The roles of gonadotrophin releasing hormone (GnRH) and a GnRH agonist (GnRHa) (r>Ala6-Met-Leu7-Pro-iV-ethylamide) in controlling pulsatile human chorionic gonadotrophin (HCG) secretion by superfused placental explants in the first trimester were examined. One minute pulses of both GnRH and GnRHa had a biphask effect upon pulsatile HCG secretion. GnRHa was maximally effective at 10" l0 M concentration, at 10" " M the effect was mild while at 10~8 M, no effect was noted. GnRH exerted a maximal stimulatory effect at 10~8 M; at 10"10 M no effect was seen, while at 10~7 M the effect was mildly stimulatory. This was evaluated by carrying out both a between and within channel type of analysis. The effect of a GnRH antagonist GnRH(ant) upon GnRH and GnRHa-induced HCG secretion was examined. Explants were incubated overnight with 10"8 M GnRH(ant), which was also continuously administered during superfusion. The addition of 1-min pulses of GnRH and GnRHa during the exposure to GnRH (ant) failed to stimulate pulsatile HCG secretion. This effect was reversible since the response to GnRH was restored within 10 min after stopping GnRH(ant) administration. In addition, by the third cycle, co-administration of GnRH(ant) for 2 min together with 10"10 M GnRHa for 1 min completely blocked the GnRHainduced effect. Continuous administration of 10"8 M GnRH(ant) decreased spontaneous HCG pulse amplitude and the area under the curve but failed to modify pulse frequency. In conclusion, GnRH appears to exert a receptor-dependent stimulatory effect upon pulsatile HCG secretion in superfusion in the first trimester placenta. Also, GnRH(ant) may reduce spontaneous HCG pulsatility by blocking endogenous GnRH action. Key words: culture/gonadotrophin releasing hormone/human chorionic gonadotrophin/placenta/superfusion

© Oxford University Press

Table I. Effect of 1-min pulses of GnRHiI upon the meari amplitude of pulsatile HCG •iecretkm by superfused placental explants i n the first trimester B

A Control GnRH(a) 1 0 - " M GnRH(a) 1 0 - 1 0 M GnRH(a) io- 9 M GnRH(a) io- 8 M

89 54 151 150 63

± ± ± ± ±

20 5 31 22 15

68 37 338 149 86

D

C ± ± ± ± ±

10 6 120* 31 12

111 91 223 186 160

± ± ± ± ±

20 24 27* 43 30

80 87 300 146 121

± ± ± ± ±

14 22 71* 17 25

Point A represents the value in mlU of HCG (mean ± SE per mg protein) at the time that the GnRH pulse was given. Points B, C and D represent the value of HCG mill (mean ± SE per mg protein) at 6, 12 and 18 min after the pulse was given. Data are the average of results obtained in three different placentas, each done in duplicate channels. In each placenta the effect of the entire dose range of GnRHa was determined. The collection interval was 6 min. *P < 0 05 versus control channel and 1 0 " " M and 10" 8 M concentrations.

1063

E.R.Barnea, M.Kaplan and Z.Naor

O •

O CONTROL • GnRHa

tein

minutes

o

1

400350-

i

300-

/

250200-

100-

o

1

\

/

\

/

\

*

150-

•

o

1

*

OH

s

1

50-

' K/

A

KJ \J^m

(-H

00

w

60

II30

"

90

GnRH a 10"9 M

GnKHa 10~10M

I

1

120

150

GnRHa 10-B M

Fig. 1. Effect of 1-min pulses of various concentrations of GnRHa on pulsatile HCG secretion. *P < 0.01 versus control channel.

Table II. Effect of 1-min pulses of GnRHa upon pulsatile HCG secretion by superfused explants as calculated by the area under the curve Treatment GnRH(a) GnRH(a) GnRH(a) GnRH(a)

Area (% ± SEM) 10"" M 10" l 0 M 1CT9 M 10" 8 M

+5% + 161% +42% + 7%

±09 ± 5.1* ±23 ± 1.2

The data show the average results obtained in three different placentas each studied with duplicate channels. The results denote the percentage change over control channel values. In each placenta, the effect of the entire dose range of GnRHa was determined. *P < 0.05 versus control channels and other GnRHa concentrations

plateau and a decline until term), it is important to establish the role of GnRH upon HCG secretion at this critical early stage. Recently, we reported that superfused first trimester placenta! explants secrete HCG in a pulsatile manner, which were further stimulated by pulses of gonadotrophin releasing hormone agonist (GnRHa) (Barnea and Kaplan, 1989) and epidermal growth factor (Barnea et al., 1990). In those reports, using a 2.4-min collection interval, we detected short-term pulsatility. Placental GnRH has not been fully characterized and therefore it is defined as GnRHlike material (Petraglia et al., 1987). The properties of the GnRH receptor in first trimester placenta are not known. Since GnRH analogues are long-acting drugs, inadvertant exposure to them may take place after implantation, which makes an examination of the effect of GnRHa on the early placenta especially important. 1064

In the present study, we have characterized the superfusion system and have examined the effects of GnRH, GnRHa and the inhibitory effect of a gonadotrophin releasing hormone antagonist, (GnRH(ant)), upon decapeptide-induced pulsatile HCG secretion. Here we show that both GnRH and its agonist stimulate HCG secretion through a receptor-dependent mechanism. Materials and methods GnRH, GnRH agonist (r>Ala6-Met-Leu7-Pro9-Af-ethylamide) and antagonist (D-pGlu'.pClPhe^D-Trp3-6) were received as gifts from Dr R.Millar (USA). Dulbecco's modified Eagle's media (DMEM) was purchased from Beit Ha-Emek, Israel. HEPES was purchased from Sigma Chemicals, St Louis, USA. Preparation of placentas A total of 30 8-11 week old placentas were studied. After obtaining informed consent, elective pregnancy terminations by vacuum curettage were performed at the Rambam Medical Centre. Following the procedure, placental tissue was immersed in 0.9% NaCl solution and carried aseptically to the laboratory. Small placental fragments of 100-150 mg were dissected out in a sterile fashion. Subsequently the explants were rinsed several times in DMEM containing antibiotic solution composed of 5000 IU penicillin, 50 /tg streptomycin and 10 000IU Fungizone (amphotericin B) per ml. Then the explants were transferred to the superfusion system described below.

GnRH and gonadotrophin releasing hormone agonist effect on placental HCG secretion

Table m . Effect of 1-min pulses of GnRH upon pulsatile HCG secretion as calculated by the area under the curve Area (% ± SEM)

Treatment GnRH GnRH GnRH GnRH

10" l 0 M 1CT9 M 1CT8 M 10" 7 M

+2% +9% + 110% + 28%

± ± ± ±

0.1 1.1 9.6' 2.3

The data represent the average of the results obtained in three different placentas run in duplicate channels The results denote the percentage change over values found in control channels. In each placenta, the effect of the entire dose range of GnRH was determined *P < 0.05 versus control channel and other GnRH concentrations.

Table IV. Effect of 1-min pulses of GnRHa or GnRH upon pulsatile HCG secretion in superfusion. The results were calculated by the area under the curve compared with values obtained prior to administration of the pulse within the same channel Mean ± SEM of three placentas run in duplicate channels Initial area before pulse

Area after pulse

GnRHa 1 0 " " M io-'° M 10-9 M io-8 M

81 137 72 57

86 1069 372 65

± ± ± ±

11 150 22* 14

GnRH 10" 10 M 10- 9 M 10" 8 M 10~7 M

81 ± 24 71 ± 22 150 ± 12 92 ± 27

89 85 387 167

± ± ± ±

21 19 45* 11*

± ± ± ±

23 32 12 12

*P < 0.05 versus initial area.

Superfusion culture A superfusion apparatus (Acusyst, Endotronics, St Paul, MN) with a multichannel peristaltic pump and fraction collector (model 272, ISCO, Durham, NC) were used to study the short-term dynamics of HCG secretion. The Acusyst system was a small volume perfusion system which consisted of four major elements: a multichannel peristaltic pump, a system of tubing sets, a heat gas exchanger and six micro-chambers each of 1 ml volume. The peristaltic pump drew fluid from the medium and delivered it through the tubing sets. The medium was carried through the heat gas exchanger, where it was warmed and brought to equilibrium with the desired gas. From the heat gas exchanger the medium was delivered to the micro-chambers to supply the tissue being cultured. Output from the micro-chambers was collected in a fraction collector for analysis. We added to the system a second peristaltic pump that permitted us to introduce different agents at a controlled rate. The explants were placed into the culture chambers and a HEPES (18 mM)-DMEM solution was washed through in an atmosphere of 5 % CO2 and 95% air at 37°C. Experiments were conducted for a period of 120 — 240 min; a sample from the effluent was collected every 2.4—6 min at 1 ml/sample. In each experiment, one channel served as control and four served as experimental channels. At given intervals, a pulse of test agent was given through the second peristaltic pump equipped with a digital fiowmeter (ISMATECH DD, Chicago, IL). This superfusion system was validated using several control

experiments. First, we found that addition of HEPES to the medium enabled us to keep the pH constant through the experiments (7.42 ± 0.02). We also monitored pCC>2 and pOj during experiments and found them constant (pCC>2 = 34 mm Hg, pO2 = 160 mm Hg). Second, we infused albumin continuously and collected the effluent in order to eliminate the possibility that the superfusion apparatus causes the HCG pulsatility through an artefact. We found that the albumin concentration remained constant in the effluent fractions collected every 2.4 min. Third, we found that pulses of media alone did not modify the pattern of spontaneous pulsatile HCG secretion. Fourth, we measured the length of time that was required for a 1-min albumin pulse to clear completely from the superfusion: within 12 min all infused albumin was cleared. Finally, we examined tissue viability following superfusion using vital staining: explants stained with try pan blue and observed under a light microscope for morphological changes using standard criteria appeared to be viable. Radioimmunoassay Measurements of HCG in the media were made using standard radioimmunoassay techniques (Barnea and Kaplan, 1989; Bamea et al., 1990). The samples were reacted with three monoclonal antibodies to HCG. Two antibodies were labelled with I25I and one with fluorescein, and they all attached to unique sites on the beta subunit of the HCG molecule. Sheep antiserum to fluorescein was also added in excess. This rapidly and specifically bound to the HCG —monoclonal antibody complex and was sedimented by a magnetic separator. A simple washing step reduced nonspecific binding to a minimum for increased low-end precision. For a given placenta, all samples were run in a single assay. Intraassay variability for HCG was 1.7%. In general, no dilution of the media was required to measure HCG. Additional control recovery studies were performed by adding purified intact HCG to pooled media samples. HCG values were determined before and after such an addition, and the percentage recovery of added HCG was calculated and found to be between 94 and 100%. Data for HCG was expressed as mlU/mg protein. The protein content of the placenta was measured by the method of Lowry et al. (1951). Statistics Statistical analysis was performed using Student's /-test and by determining the area under the curve using Simpson's rule on a personal computer. A probability of P < 0.05 was considered statistically significant. The figures are representative of data generated in at least three different placentas, each run in duplicate channels. Data given in the tables are of mean ± SEM and statistical analysis of composite data generated in these experiments. Pulse pattern analysis was carried out using the PULSAR program obtained as a gift from G.R.Merriam, NIH. Results Effect of GnRHa Addition of 1-min pulses of 10"'° M GnRHa caused a marked 4- to 6-fold increase in HCG secretion compared to control channels (data not shown). The effect of 1 min pulses of 1065

E.R.Barnea, M.Kaplan and Z.Naor

- CONTROL •• TREATMENT

A

GnRH antagonist 10~8M

n

•i—i

CD -1-3

o OH

O

0

40

80

120

10

160

minutes

: GnRHa 10- M -o CONTROL - ° TREATMENT

B

GnRH antagonist 10

o

a 200 0

40 80 8 : GnRH 10~ M 10

120"

160

minutes 8

Fig. 2. A Inhibition of action of 1-min pulses of 10~ M on HCG secretion by continuous 10~ M GnRH(ant) administration. B Inhibition of effect of 1-min pulses of 10~8 M GnRH on HCG secretion by continuous 10^8 M GnRH(ant) administration. The subsequent 10~8 M GnRH pulse was effective within 10 min after stopping the antagonist administration (B) also shows in detail the pulsatile panem of HCG secretion in the control channel (see the different scale).

1066

GnRH and gonadotropbin releasing hormone agonist effect on placenta) HCG secretion

o •

o CONTROL • TREATMENT || : 2 min GnRHant 10~8M+ 1 min GnKHa 10"10M

CD O

s + 10 0

80

minutes Fig. 3. Short-term inhibitory effect of 10"8 M GnRH(ant) on GnRHa-induced effect. By the third cycle of administration the effect was completely blunted. *P < 0.01, °P < 0.05 versus control channel. The lower set of data shows in detail the pulsatile pattern of HCG secretion in the control channel (note the different scale).

10 " - 1 0 8 M GnRHa was evaluated. Table I shows that mean ± SEM pulse amplitude increased from 10~" to 10"10 M. At 10~ 9 -10" 8 M concentrations, the pulse amplitude decreased compared with that noted at 10~'°M (Table I, Figure 1). Table II illustrates the significant changes in the calculated area under the curve (mean ± SEM) observed following administration of 1 0 " " - 1 0 " 8 M of GnRHa for 1 min. Effect of GnRH One minute pulses of 10" 8 M GnRH had a significant stimulatory effect upon first trimester placenta! explants in superfusion compared to controls. The effect of 10~9 M was not significant. At lower (10~10 M) and higher concentrations (10~7 M) the effect was also less effective (data not shown). Table IE shows the changes in the calculated area under the curve (mean ± SEM) following GnRH exposure (P < 0.05). In order to evaluate further the effect of GnRH and GnRHa, a within-channel type of analysis was also carried out. Table IV compares the initial area under the curve prior to the 1-min pulse and the area of the peak following it within the same channel, thus confirming that the maximal effectiveness was observed at 10" l0 M for GnRHa and at 10"8 M for GnRH. Inhibitory effect of GnRH(ant) GnRH(ant) completely blocked GnRH- and GnRHa-induced HCG secretion in superfused first trimester placenta! explants.

Table V. Pulse characteristics of peak HCG secretion by placental explants superfused with 10" 8 M GnRH(ant) for 90 min as analysed by PULSAR AUC Frequency Amplitude Control 147 GnRH(ant) 79

17.2 18.7

Height

Length

Area

7.1 ± 1.8 7.8 ± 1 7 0.08 ± 0.01 8.1 ± 2.6 4.4 ± 1.1*8.1 ± 1.5 0 09 ± 0.01 6 1 ± 2 . 6

Data are of mean ± SEM of three placentas run in duplicate. AUC, area under the curve. *P < 0 05 versus control channel

An overnight incubation with 10 8 M GnRH(ant) followed by continuous exposure to the antagonist were required to block the effect of a 1-min 10~8 M GnRH pulse (Figure 2A). This was reversible since upon stopping GnRH(ant) administration, within 10 min the response to a 1-min pulse of GnRH was restored (Figure 2B), while the response to a pulse of GnRHa given for 1-min was not restored. Co-administration of 10"8 M GnRH(ant) for 2 min together with 10" l0 M GnRHa for 1-min caused a progressive decrease in the amplitude of GnRHa-induced HCG secretion (P < 0.05). The GnRHa effect was completely blocked by the third cycle of administration (Figure 3). Continuous administration of 10" 8 M GnRH(ant) decreased the area under the curve and the peak amplitude of spontaneous pulsatile HCG secretion, as analysed by the PULSAR program (Table V). 1067

E.R.Barnea, M.Kaplan and Z.Naor

Discussion A role for GnRH in controlling placental HCG secretion has been suggested previously (Siler-Khodr and Khodr, 1979; Belisle et al., 1984). However, no firm evidence has been presented thus far to support such a view. In this report, we provide evidence that supports the role of GnRH in controlling HCG secretion by the first trimester placenta. The stimulatory effect was biphasic and was obtained with pulses of both GnRH and its potent agonist. The effect appeared to be receptor-dependent since GnRH(ant) blocked both the GnRH- and GnRHa-induced effects. In the case of GnRH, this effect was reversible. Moreover, GnRH(ant) reduced the amplitude of spontaneous HCG pulses, the area under the curve but not the frequency, suggesting that the effect of endogenous GnRH is blocked by the compound. The stimulatory effect of GnRH pulses was documented by changes in the amplitude of different points within a peak and the resulting area under the curve, compared to the control channel. A concentration of 10~8 M GnRH was most effective while that for the agonist was 10~10 M, which corresponds well with the 100-fold difference in their potency (R.Millar, personal communication). The maximally effective dose of GnRH is similar to the placenta! GnRH content, and to the affinity constant (Ka) of the placental high-affinity GnRH receptor (Siler-Khodr and Khodr, 1979; Belisle et al., 1984). We found that the GnRH effect was more pronounced when given in short pulses since in long-term (24 h) static cultures the effect was only mild (Khodr and Siler-Khodr, 1978; Kim et al., 1987; Barnea et al., unpublished). Thus, our superfusion model enables a noticeable expression of GnRH effect without desensitizing the local receptors, which may occur in static cultures. A similar down-regulation was probably produced by high doses of 10~8 M GnRHa given for 1 min (Tables I and IT). This may be similar to previously reported results in rat pituitary cultures (Loumaye and Catt, 1983). Both the amplitude and frequency of HCG pulses increased following pulsatile GnRHa administration (Barnea and Kaplan, 1989). The GnRH effect was produced within 6 min which indicates that the stored HCG was released from local storage granules (Morrish et al., 1987). The rapidity of the GnRH effect resembles that seen in the pituitary (Loumaye and Catt, 1983). The contact between the GnRH and the placenta lasted only a few minutes, as seen by the 1 -min pulse of albumin which was completely cleared within 12 min. However, the effect of GnRH upon HCG secretion lasted for nearly 30 min, suggesting that GnRH binds locally to produce its prolonged action. Specific high-affinity GnRH sites in the young placenta have been detected by us recently (Barnea et al., unpublished). Maximally effective doses of GnRH and GnRHa were blocked by GnRH(ant), reversibly in the former case (Figure 2). The lack of GnRHa effect in this case may lie in the delicate but thus far not fully elucidated differences between the peptide responses. Perhaps the close structural similarity of GnRH to the native compound enables the antagonist effect to be overcome rapidly. Moreover, as seen in Figure 3, 2-min pulses of GnRH (ant) combined with GnRHa for 1 min caused a progressive desensitization. However, this was not noted with the agonist alone (Figure 2). 1068

The GnRH(ant) blocking effect in long-term incubations was not due to its toxicity since it was also produced following shortterm application (6 min) (Figure 3). Further evidence of this is given by the promptly restored response to GnRH after the antagonist was discontinued. Continuous administration of GnRH(ant) decreased both the area under the curve and the peak amplitude of spontaneous HCG pulsatility but did not modify the pulse frequency. This provides indirect evidence that the antagonist inhibited the endogenous peptide, thus reinforcing the physiological role of endogenous GnRH in controlling the amplitude of HCG secretion. This also suggests that the spontaneous pulse frequency is not controlled by a paracrine/autocrine effect of GnRH. Indeed, we found detectable levels of GnRH-like material in the superfusion effluent (Barnea et al., unpublished). Similar to reports of work in the pituitary gland (Kenisberg etal., 1984), high doses (10~6 M - 1 0 " 7 M) of GnRH(ant) increased HCG secretion in superfusion (Barnea et al., unpublished). The present data and methodology provide a useful model for initiating investigations into mechanisms of GnRH action on the placenta, which are currently being carried out in our laboratory. In our view, the present data support the role of endogenous GnRH in regulating HCG secretion, through an autocrine/paracrine mechanism. This role is similar to the endocrine role exerted by GnRH on gonadotrophin secretion in the pituitary. The presence of a local placenta! neurotransmitter—GnRH—HCG axis is further supported by the recent findings that placental GnRHlike material secretion can also be modulated by various hormones (Petraglia etal., 1987). Acknowledgements This work is supported in part by a grant from Ministry of Health Chief Scientist and the Juvenile Diabetes Foundation to E.R.B.

References Barnea,E.R. and Kaplan,M. (1989) Gonadotropin releasing hormoneinduced and progesterone-inhibited pulsatile secretion of human chorionic gonadotropin in the first trimester placenta in vitro. J. Clin. Endocrinol. Metab., 69, 215-218. Bamea,E.R., Feldman.D., Kaplan,M. and Mornsh.D.W. (1990) The dual effect of epidermal growth factor upon human chorionic gonadotropin secretion by the first trimester placenta in vitro. J. Clin. Endocrinol. Metab., 71, 923-928. Belisle,S., Guevin,J.F., Bellabarba.D. and Lehoux.J.G. (1984) Luteinizing hormone-releasing hormone binds to enriched human placental membranes and stimulates in vitro the synthesis of bioactive human chorionic gonadotropin. J. Clin. Endocrinol. Metab., 59, 119-124. Currie.A.J., Fraser.H.M. and Scharpe.R.M. (1981) Human placental receptors for luteinizing hormone releasing factor. Biochem. Biophys. Res. Commun., 99, 332-336. Kenisberg,D., Littman.B.A. and Hodgen.G.D. (1984) Medical hypophysectomy I: Dose-response using a GnRH antagonist. Fertil. Steril, 42, 112-116. Khodr,G.S. and Siler-Khodr,T.M. (1978) The effect of luteinizing hormone-releasing factor on human chorionic gonadotropin secretion. Fertil. Steril., 30. 3O1-3W. Kim.S.J.. Namkoong,S.R.. Lee.J.W.. Jung.J.K.. Kang,B.C. and

GnRH and gonadotrophin releasing hormone agonist effect on placental HCG secretion Park.J.S. (1987) Response of human chorionic gonadotropin to luteinizing hormone-releasing hormone stimulation in the culture media of normal human placenta, choriocarcinoma cell lines, and in the serum of patients with gestational trophoblastic disease. Placenta, 8, 257-263. Loumaye,E. and Catt.K.H. (1983) Agonist-induced regulation of pituitary receptors for gonadotropin releasing hormone. J. Biol. Chem., 258, 12002-12004. Lowry.O.H., Rosebrough,N.A., Farr.A.L. and Randall,R.J. (1951) Protein measurement with the folin phenol reagent. J. Biol. Chem., 193, 265-230. Morrish.D.W., Marusyk.H. and Siy.O. (1987) Demonstration of specific secretory granules for human chorionic gonadotropin in placenta. J. Histochem. Cytochem., 35, 93-101. Nekda,M.V., Horwath,A., Ge,L.J., Coy,D.H. and Schally.A.V. (1982) Suppression of ovulation in the rat by an orally active antagonist of luteinizing hormone-releasing hormone. Science, 218, 160—165. Owens,O.M., Ryan,K.J. and Tulchinsky,D. (1981) Episodic secretion of human chorionic gonadotropin in early pregnancy. J. Clin. Endocrinol. Metab., 53, 1307-1310. Padmanabhan.V., Sonstein,J., Olton,P.L., Nippold.T., Menon,K.M.J., Marshall,J.C, Kelch,R.P. and Beitins.I.Z. (1989) Serum bioactive follicle-stimulating hormone-like activity increases during pregnancy. J. Clin. Endocrinol. Metab., 69, 968-973. Petraglia.F., Lim.A.T.W. and Vale.W. (1987) Adenosine 3',5'-monophosphate, prostaglandins and epinephrine stimulate the secretion of lmmunoreactive gonadotropin-releasing hormone from cultured human placental cells. J. Clin. Endocrinol. Metab., 65, 1020-1024. Pohl.C.R., Richardson,D.A., Hutchison,J.S., Germak.J.A. and Knobil,E. (1984) Hypophysiotropic signal frequency and the functioning of the pituitary ovarian system in the rhesus monkey. Neuroendocrinology, 39, 256—263. Popkin.R., Bramley.T.A., Currie.A., Shaw.R.W., Baird,D.T. and Fraser.H.M. (1983) Specific binding of luteinizing hormone releasing hormone to human luteal tissue. Biochem. Biophys. Res. Commun., 114, 750-754. Siler-Khodr,T.M. and Khodr,G.S. (1979) Extrahypothalamic luteinizing hormone releasing factor (LRF): release of immunoreactive LRF in vitro. Fertil. Sterii, 39, 294-299. Siler-Khodr,T.M , Khodr.G.S., Valenzuela.G and Rhode,J. (1986) GnRH effects on placental hormones during gestation: II Progesterone, estrone, estradiol and estriol. Biol. Reprod., 34, 255—263. Received on February 8, 1991; accepted on May 21, 1991

1069