Autocrine Regulation of Gonadotropin-Releasing Hormone Secretion ...

0013-7227/99/$03.00/0 Endocrinology Copyright © 1999 by The Endocrine Society

Vol. 140, No. 3 Printed in U.S.A.

Autocrine Regulation of Gonadotropin-Releasing Hormone Secretion in Cultured Hypothalamic Neurons LAZAR Z. KRSMANOVIC, ANTONIO J. MARTINEZ-FUENTES, KRISHAN K. ARORA, NADIA MORES, CARLOS E. NAVARRO, HAO-CHIA CHEN, STANKO S. STOJILKOVIC, AND KEVIN J. CATT Endocrinology and Reproduction Research Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892 ABSTRACT Episodic hormone secretion is a characteristic feature of the hypothalamo-pituitary-gonadal system, in which the profile of gonadotropin release from pituitary gonadotrophs reflects the pulsatile secretory activity of GnRH-producing neurons in the hypothalamus. Pulsatile release of GnRH is also evident in vitro during perifusion of immortalized GnRH neurons (GT1–7 cells) and cultured fetal hypothalamic cells, which continue to produce bioactive GnRH for up to 2 months. Such cultures, as well as hypothalamic tissue from adult rats, express GnRH receptors as evidenced by the presence of high-affinity GnRH binding sites and GnRH receptor transcripts. Furthermore, individual GnRH neurons coexpress GnRH and GnRH receptors as revealed by double immunostaining of hypothalamic cultures. In static cultures of hypothalamic neurons and GT1–7 cells, treatment with the GnRH receptor antagonist, [D-pGlu1, D-Phe2, D-Trp3,6]GnRH

T

caused a prominent increase in GnRH release. In perifused hypothalamic cells and GT1–7 cells, treatment with the GnRH receptor agonist, des-Gly10-[D-Ala6]GnRH N-ethylamide, reduced the frequency and increased the amplitude of pulsatile GnRH release, as previously observed in GT1–7 cells. In contrast, exposure to the GnRH antagonist analogs abolished pulsatile secretion and caused a sustained and progressive increase in GnRH release. These findings have demonstrated that GnRH receptors are expressed in hypothalamic GnRH neurons, and that receptor activation is required for pulsatile GnRH release in vitro. The effects of GnRH agonist and antagonist analogs on neuropeptide release are consistent with the operation of an ultrashort-loop autocrine feedback mechanism that exerts both positive and negative actions that are necessary for the integrated control of GnRH secretion from the hypothalamus. (Endocrinology 140: 1423– 1431, 1999)

HE HIERARCHICAL secretion of GnRH and gonadotropins (LH and FSH) is regulated by gonadal steroids, which exert positive and negative actions at the hypothalamus and the pituitary gland (1). The increase in gonadotropin secretion following gonadectomy stabilizes at a level that is determined by short- and ultrashort-loop feedback between the pituitary gland and the hypothalamus (2– 4). Autocrine regulation of neuropeptide release through an ultrashort negative feedback mechanism was first suggested in studies on the control of FSH-releasing factor (FSH-RF) secretion (5). More recently, in vitro studies employing hypothalamic slices, cultured hypothalamic cells, and immortalized GnRH neurons (GT1 cells) have been applied to the analysis of GnRH secretion and its regulation (6 –11). Electrophysiological studies on GnRH-producing cell lines (12–14) and GnRH-containing embryonic neurons (15) have demonstrated the expression of a wide variety of voltage and ligand-gated channels. Tetrodotoxin-sensitive sodium current, as well as Ca21 current with transient and sustained dihydropyridine-sensitive components, have been observed (13, 14). GnRH-producing neuronal cell lines also exhibit episodes of depolarizing electrical activity and fluctuations in [Ca21]i (16, 17). The presence of connexin 26 proteins in GT1–7 cells (18) suggests that gap junction coupling between

GnRH-producing neurons could serve to coordinate their pulsatile secretion. In rhesus monkeys, in vivo measurements of multiunit electrical activities (MUA) from the medialbasal hypothalamus have shown invariant synchrony between abrupt increases in frequency of MUA volleys and the initiation of LH pulses (19). Although it is not clear whether this electrical activity originates from GnRH cells or other neuronal elements, it is probable that GnRH neurons with intrinsic electrical activity participate in the formation of MUA. The ability of GT1 cells to exhibit episodic GnRH release in the absence of other cell types indicates that intrinsic factors, such as autocrine regulation of neurosecretion, could be important determinants of pulsatile GnRH release. The finding that GT1–7 cells express GnRH receptors, agonist activation of which influences the pattern of pulsatile GnRH release by changing pulse frequency and amplitude (20), is consistent with this proposal. The present studies were performed to determine whether such autocrine regulation through endogenous GnRH receptors is also operative in normal GnRH neurons. For this purpose, cultured fetal hypothalamic cells were employed to analyze the expression of GnRH receptors, and the influence of GnRH agonist and antagonist analogs on the dynamics of GnRH release from native GnRH neurons.

Received May 18, 1998. Address all correspondence and requests for reprints to: Kevin J. Catt, M.D., Ph.D., Endocrinology and Reproduction Research Branch, Building 49, Room 6A-36, NICHD, NIH, Bethesda, Maryland 20892. E-mail: [email protected].

Materials

Materials and Methods Hypothalamic tissue was removed from fetuses of 17-day pregnant Sprague Dawley rats. The borders of the excised hypothalami were delineated by the anterior margin of the optic chiasm, the posterior

1423

1424

AUTOCRINE REGULATION OF GnRH SECRETION

Endo • 1999 Vol 140 • No 3

margin of the mammillary bodies, and laterally by the hypothalamic sulci. After dissection, hypothalami were placed in ice-cold dissociation buffer containing 137 mm NaCl, 5 mm KCl, 0.7 mm Na2HPO4, 25 mm HEPES, 100 mg/liter gentamicin, pH 7.4. The tissues were washed and then incubated in a sterile flask with dissociation buffer supplemented with 0.2% collagenase (activity 149 U/mg; Worthington, Freehold, NJ), 0.4% BSA, 0.2% glucose, and 0.02% of DNase I (Sigma Chemical Co., St. Louis, MO). After 60 min of incubation in a 37 C water bath with shaking at 60 cycles/min, the tissue was gently triturated by repeated aspiration into a smooth-tipped Pasteur pipette. Incubation was continued for another 30 min, after which the tissue was again dispersed. The cell suspension was passed through sterile mesh (200 mm) into a 50-ml tube, sedimented by centrifugation for 10 min at 200 3 g, and washed once in dissociation buffer and once in culture medium consisting of 500 ml DMEM containing 0.584 g/liter l-glutamate and 4.5 g/liter glucose (Sigma Chemical Co.), mixed with 500 ml F-12 medium containing 0.146 g/liter l-glutamine, 1.8 g/liter glucose (Sigma Chemical Co.), 100 mg/ml gentamicin, 2.5 g/liter sodium bicarbonate, and 10% heat-inactivated FCS (Gibco BRL-Life Technologies, Gaithersburg, MD). Each dispersed hypothalamus yielded about 1.5 3 106 cells.

1 m NaOH containing 0.1% SDS and analyzed for bound radioactivity in a g-spectrometer. GnRH receptors were also analyzed in membrane fractions prepared from cultured fetal hypothalamic cells. After 10 days in culture, cells were washed twice with binding buffer containing 10 mm (Tris[hydroxymethyl]aminomethane)-Trizma base, 0.1% BSA, 1 mm DL-dithiothreitol, 10 mg/ml aprotinin, 0.1 mm phenylmethylsulfonyl fluoride, and 1 mm EDTA in PBS, pH 7.4. The cells were broken in a glass homogenizer, briefly sonicated, and sedimented at 400 mg for 15 min. The resulting pellets were centrifuged at 10,000 3 g for 30 min and resuspended in binding buffer. For the assay procedure, 50 –100 mg receptor protein was incubated with 150 pm 125I-labeled [d-Ala6]Ag for 60 min on ice in the presence of increasing concentration of the test peptides. Nonspecific binding was assessed in the presence of 1 mm of [d-Ala6]Ag. Incubations were performed in 12 3 75-mm borosilicate glass tubes in a total volume of 0.5 ml and terminated by dilution with 4 ml ice-cold PBS (pH 7.4), followed by filtration through glass fiber filters (GF/C Whatman Daigger Scientific, Wheeling, IL) in a multiple holder. The filters were washed three times with 4 ml PBS, and retained radioactivity was determined in a g-spectrometer.

Perifusion procedure

Preparation of hypothalamic RNA

Dispersed hypothalamic cells were incubated in 50-ml tubes containing 1.5 3 107 cells, 0.3 ml preswollen Cytodex-2 beads (Pharmacia, Piscataway, NJ), and 30 ml of culture medium for 24 h in 5% CO2/air. The suspension was then transferred into 60-mm dishes and culture was continued for 14 – 60 days, with replenishment of culture medium every second day. Before perifusion, the cell-bead mixture was collected by sedimentation and resuspended in Krebs-Ringer buffer continuing 1 mg/ml BSA, 1 mg/ml glucose, 20 mm bacitracin, pH 7.4. After gassing for 1 h with 95% O2/5% CO2, the beads were loaded into a temperaturecontrolled 0.5 ml chamber (Endotronics, Inc.; Minneapolis, MN). Perifusion was performed at a flow rate of 10 ml/h at 37 C for at least 1 h to establish a stable baseline before addition of agents made up in the same medium. Fractions were collected at either 1- or 5-min intervals and stored at 220 C before RIA using 125I-GnRH (Amersham, Arlington Heights, IL), GnRH standards (Peninsula, Belmont, CA), and primary antibody donated by Dr. V. D. Ramirez, University of Illinois (Urbana, IL) (21). The intraassay and interassay coefficients of variation at 80% binding in standard samples (15 pg/ml) were 12–14%, respectively.

Total RNA was prepared from hypothalamic tissue and cells by the acid guanidinium thiocyanate-phenol-chloroform method as described by Chomczynski and Sacchi (23). Hypothalami were removed from adult rats by dissection and rapidly frozen on dry ice, and 10-day cultures of fetal rat hypothalamic cells were washed with ice-cold PBS and frozen on dry ice. RNA concentrations and purity were determined spectrophotometrically at 260 and 280 nm (24).

Static culture of hypothalamic cells and GT1–7 cells Cells were cultured in 12-well plates [2 3 106 cells per well for hypothalamic cells and 1 3 106 cells per well for GT1–7 cells, provided by Dr. R. I. Weiner, University of California San Francisco (San Francisco, CA) (22)] in 2 ml of culture medium. GnRH agonist (des-Gly10-[dAla6]GnRH N-ethylamide; [d-Ala6]Ag, 50 pm) and antagonist ([d-pGlu1, d-Phe2, d-Trp3,6]GnRH; [d-pGlu]Antag, 50 pm) analogs were added to the culture wells immediately after plating the cells. In some experiments, an additional GnRH antagonist analog ([Ac-d-Nal (2)1, d-Phe(pCl)2, d-Pal (3)3, d-Cit6, d-Ala10]GnRH; SB-75, donated by Dr. A. V. Schally, VA Hospital, New Orleans, LA) was used. Initial levels of GnRH (time “0”) in control and treated groups were measured immediately after plating the cells and adding the peptides. GnRH production was measured at 24-h intervals after removing 1 ml of medium and replacing 1 ml of fresh medium, and was calculated as cumulative production over 8 to 9 days of culture. The GnRH content of cultured hypothalamic cells and GT1–7 neurons was measured by RIA after brief sonication in 1 m Na2CO3, followed by adjustment of the pH of samples to 7.5.

Radioligand binding assays Plasma membrane receptors for GnRH (GnRH-R) were analyzed by binding studies with 125I-des-Gly10-[d-Ala6]GnRH N-ethylamide (Hazleton, VA). The radioligand (150 pm) and nonradioactive peptides were added in 100-ml aliquots to 1-week-old monolayer cultures of hypothalamic cells maintained in 12-well Falcon plates. After incubation to equilibrium for 60 min at room temperature, the cells were washed three times with ice-cold PBS containing 0.1% BSA, then solubilized in

RT-PCR RT-PCR was carried out using the GeneAmp Thermostable rTth reverse transcriptase RNA PCR kit (Perkin Elmer, Norwalk, CT). First strand complementary DNA (cDNA) was synthesized using total RNA and random primers (Invitrogen, San Diego, CA) and PCR amplification was performed with gene-specific primers based on sequences in the transmembrane domains of the mouse pituitary GnRH receptor (25). The sequence of the primers used for PCR were sense (1S) 59-GTGACCGTGACTTTCTTC-39; and antisense (7AS) 59-GTCGAAGCACGGGTTTAG-39. The primers used for nested PCR were sense (3S) 59-CTCAGCTATCTGAAGCTCTTC-39; and antisense (6AS) 59-GACGACAAAGGAGGTAGCG-39. The numbers in parentheses refer to the transmembrane domains. Nested PCR was performed in combination with primers 1S and 6AS, or 3S and 6AS. The reaction conditions for amplification were 94 C for 1 min, 45 C for 1 min, and 72 C for 1 min for 35 cycles, and were carried out in a Perkin Elmer Cetus DNA thermal cycler. Twenty-microliter aliquots of the 100 ml PCR reaction mixture were then electrophoresed on a 1.0% agarose gel and visualized by staining with ethidium bromide (24). A control without the reverse transcriptase step was performed to exclude the possibility of contamination with genomic DNA. The authenticity of the PCR-amplified products was confirmed by Southern blot analysis using a32P-labeled GnRH receptor cDNA probe (25), and a final stringent wash with 0.53 saline-sodiumcitrate (SSC) and 0.1% SDS at 55 C.

Immunocytochemical determination of GnRH and GnRH receptor in cultured hypothalamic neurons For immunocytochemical localization of GnRH and GnRH-R in cultured fetal hypothalamic cells, the enzymatically dispersed cells were plated on glass chamber slides in standard culture medium at a density of 105 cells/well. After 3 days, the medium was supplemented with 5-fluoro-2-deoxyuridine (80 mm) and culture was continued for 3 days. Before immunostaining the culture medium was removed and the slides were washed with 0.01 m PBS, fixed with Bouin’s fluid for 30 min, washed, dehydrated, and kept dry at 270 C. Immunostaining for GnRH and GnRH-R was performed by the avidin-biotin peroxidase and alkaline phosphatase methods, respectively. After hydration the cells were treated with 3% H2O2, rinsed, blocked by incubation in 10% normal goat serum in PBS, washed, and incubated overnight at 4 C with an antiGnRH polyclonal antibody (1:1000; generously provided by Dr. V. D. Ramirez). On day 2, the slides were rinsed and incubated in goat an-


1425

GnRH pulses were identified, and their parameters were determined by computerized cluster analysis (26). Individual point standard deviations were calculated using a power function variance model from the experimental duplicates. A 2 3 2 cluster configuration and a t statistic of 2 for the upstroke and downstroke were used to maintain falsepositive and false-negative error rates below 10%. The statistical significance of the pulse parameters was tested by using one-way ANOVA.

receptor cDNA probes further confirmed the authenticity of the PCR amplified products. No such products were obtained in the absence of reverse transcribed mRNA, indicating that the RNA preparation was free of genomic DNA contamination. These results demonstrate the expression of the GnRH receptor gene in the hypothalamus, consistent with the presence of GnRH receptor sites in these cells (Fig. 1A). Immunostaining with a specific GnRH antiserum revealed that about 2% of the cultured hypothalamic cells were GnRHcontaining neurons with typical bipolar morphology. The brown reaction product characteristic of the GnRH immunocytochemical precipitate complex was distributed throughout the cytoplasm and primary processes and was absent from the nucleus. On double immunostaining with both GnRH and GnRH receptor antisera, a majority of the cells that were positive for GnRH also exhibited blue granular staining for GnRH-R. This was predominantly distributed at the plasma membrane of bipolar neurons (Fig. 1C) and in some monopolar neurons (Fig. 1D). No immunostaining for GnRH-R (blue) was detectable when GnRH positive cells (brown) were incubated in normal goat serum without primary antibody for GnRH-R, and subsequently treated with the Vector blue alkaline phosphatase reagent (Fig. 1E).

Results Expression of GnRH receptors in hypothalamic neurons

Effects of GnRH receptor agonist and antagonist analogs in static cultures

Cultured hypothalamic neurons exhibited specific, highaffinity binding of the radioiodinated GnRH agonist, desGly10[d-Ala6]GnRH N-ethylamide (125I-[d-Ala6]Ag). The binding of 125I-[d-Ala6]Ag to hypothalamic cells was inhibited by GnRH and its agonist and antagonist analogs in a dose-, time-, and temperature-dependent manner. Receptor specificity was indicated by the ability of such GnRH ligands (100 nm) to inhibit radioligand binding by up to 97% (Fig. 1A), and the lack of displacement by unrelated peptides (100 nm) including angiotensin II, TRH, oxytocin, and arginine vasopressin (not shown). The concentration-dependent inhibition of 125I-[d-Ala6]Ag binding by unlabeled GnRH, and GnRH agonist and antagonist analogs, is illustrated in Fig. 1A. The estimated IC50 values for each competition curve were 23 nm for GnRH, 1.6 nm for the [d-Ala6]Ag, and 0.2 nm for the [d-pGlu]Antag. Scatchard analysis of the data for GnRH and its agonist and antagonist analogs revealed the presence of both high and low affinity binding sites. The Kd values of the high affinity sites were 1.4 nm for GnRH, 0.9 nm for [d-Ala6]Ag, and 0.3 nm for [d-pGlu]Antag. Those of the low affinity sites were 638 nm for GnRH, 478 nm for [d-Ala6]Ag, and 91 nm for [d-pGlu]Antag (Fig. 1A). The expression of GnRH receptors in cultured fetal cells and the adult hypothalamus was also demonstrated by RTPCR. Analysis of total RNA, using gene-specific primers based on sequences in transmembrane domains (TM) I and VII of the receptor, gave the expected size fragment of 840 bp. When nested PCR was performed using primers based on sequences in TMs I, III, and VI, the expected size products with primer sets from TM I and VI, and TM III and VI, corresponding to 717 and 483 bp, respectively, were observed (see Fig. 1B). Southern hybridization using GnRH

Static cultures of hypothalamic cells contained 1074 6 83 pg per 106 cells, and released 27.4 6 1.2 pg/ml into the incubation medium (n 5 6). Depolarization with 50 mm KCl caused a significant increase in GnRH release (121 6 8 pg/ ml; P , 0.01; n 5 6) with a concomitant decrease in GnRH content to 671 6 72 pg/106 cells (P , 0.05; n 5 6; Fig. 2). Static cultures of immortalized GnRH neurons contained 1442 6 105 pg GnRH per 106 cells, and released 49 6 1.4 pg/ml into the incubation medium (n 5 6). Depolarization with 50 mm KCl increased GnRH release to 163 6 24 pg/ml (P , 0.01; n 5 6), with a concomitant decrease in GnRH content to 769 6 98 pg/ml (P , 0.05; n 5 6; Fig. 2). The role of GnRH receptor activation in neuropeptide secretion was evaluated by treatment with GnRH agonist ([d-Ala6]Ag) and antagonist (SB-75) analogs that did not cross-react in the GnRH RIA. The cumulative release of GnRH during static culture of hypothalamic cells (Fig. 3A) and GT1–7 cells (Fig. 3B) was measured under basal conditions and during treatment with low concentrations of the two analogs. At the zero time point, no immunoreactive GnRH was detectable in medium of controls and treated cells. The basal release of GnRH increased from 3.9 6 1.4 pg/ml on day 1 of culture to 17.2 6 1.8 pg/ml (P , 0.01) on day 3 in culture in GT1–7 cells, and from 7.4 6 0.8 pg/ml on day 8 of culture to 14.5 6 0.6 (P , 0.01) day 5 in culture in control hypothalamic cells. Daily addition of 50 pm [d-Ala6]Ag to the culture medium did not significantly change GnRH release from either cell type, and the cumulative GnRH profile was similar to that observed in controls. However, sustained antagonist treatment with 50 pm [dpGlu]Antag caused a prominent increase in GnRH release during the first three days of culture, followed by a mono-

tirabbit IgG-biotin conjugate (1:500) followed by avidin-biotin peroxidase complex (1:350). GnRH staining was visualized with a diaminobenzidine substrate kit for peroxidase (Vector, Burlingame, CA). After GnRH staining, cells were rinsed, blocked by incubation in 10% normal goat serum, washed, and incubated overnight with a polyclonal antiserum to the GnRH-R (1:500) at 4 C. The GnRH-R antiserum was raised in a rabbit immunized with bovine thyroglobulin conjugated with a synthetic peptide corresponding to the third extracellular loop of the mouse GnRH-R (residues 291–306). The cells were then incubated in goat antirabbit IgG-biotin conjugate (1:500, Vector, Burlingame, CA) and an avidin-biotin alkaline phosphatase complex. The GnRH-R antigenic sites were visualized using a Vector Blue Alkaline phosphatase substrate kit III. Antibody specificity was determined by treating cells with GnRH antibody preadsorbed with synthetic GnRH or without primary antibody substituted with normal goat serum for GnRH-R. In double immunostaining, no blue reaction product was formed when the GnRH-R antibody was omitted after GnRH immunostaining was completed, confirming the specificity of the alkaline phosphatase reaction for the GnRH-R.

Data analysis

1426


Endo • 1999 Vol 140 • No 3

FIG. 1. Expression of GnRH receptors in hypothalamic tissue and cultured hypothalamic cells. A, Competitive inhibition of 125I-[D-Ala6]Ag binding to the particulate fraction of cultured hypothalamic cells by the unlabeled agonist, antagonist and native GnRH. B, RTPCR of GnRH receptor mRNA extracted from adult rat hypothalamus and rat fetal hypothalamic cultures. The autoradiograph shows PCR products probed with 32P-labeled GnRH receptor cDNA. Lanes 1, 2, and 3 correspond to the primer pairs 1S and 7AS, 1S and 6AS, and 3S and 6AS, respectively. The expected sizes of the fragments are indicated by arrowheads. Similar data were obtained in three separate experiments. C, Double immunostaining of GnRH (brown) and GnRH-R (blue) in bipolar hypothalamic neurons (3 1000). D, Positive immunostaining of monopolar hypothalamic neuron for GnRH (brown) and GnRH-R (blue) (3 1000). E, Lack of GnRH-R immunostaining by the alkaline phosphatase reaction (blue) in identified GnRH neurons in the absence of GnRH-R antibody.

tonic decrease to a level comparable to that observed in untreated cells. Effects of GnRH agonist and antagonist analogs on episodic neurosecretion

Cultured fetal hypothalamic cells consistently exhibited pulsatile GnRH release when perifused in the absence of exogenous stimuli (Fig. 4A). Sample collections at 1-min intervals provided four to five data points per peak, and revealed clearly defined episodes of GnRH release. Under basal conditions, the mean amplitude of each episode of GnRH release, calculated from the five highest points for each peak, was 7.6 6 1.3 pg/ml. Transient depolarization by exposure to 35 mm KCl significantly increased the mean peak

amplitude to 12.3 6 2.4 pg/ml (P , 0.05). Over 2.5 h of perifusion, 160 data points were collected and 9 peaks were detected by cluster analysis. In this experiment, the mean interval between peaks was 15.3 6 3.0 min and the mean peak width was 14 6 3.6 min. When the GnRH secretory profile was obtained by plotting every fifth measurement from the original data set, the GnRH secretory profile was similar to that given by all data points (Fig. 4B). The number of peaks detected was the same (9/2.5 h), the mean interval between peaks was unchanged at 16.8 6 4.5 min, and there was a slight decrease in mean peak width (11.6 6 4.3 min) compared with the original data set. There was also a close correlation between the curve for GnRH release obtained by calculating the cumulative value during


FIG. 2. GnRH release (upper panel) and content (bottom panel) in static cultures of hypothalamic cells and GT1–7 neurons. GnRH secretion increased significantly (P , 0.01) and GnRH content decreased significantly (P , 0.01) in response to K1 depolarization. Statistical differences were calculated using the one factor ANOVArepeated measurement test.

5-min collections and that derived from the original data set (Fig. 4C). After this transformation, nine peaks were detected by cluster analysis, the mean interval between peaks was 16.2 6 2.3 min, and the mean peak width was 11.6 6 3.5 min. Because analysis of 5-min fractions obtained at a flow rate of 10 ml/h was adequate to monitor episodic GnRH release, these conditions were used in subsequent perifusion experiments. The biological activity of GnRH released by hypothalamic cells was evaluated by directing the perifusion medium from the hypothalamic cells into a downstream chamber containing anterior pituitary cells. Basal LH release was low before connection with a chamber containing hypothalamic cells (Fig. 4D, open circles) and increased significantly after connection was established (closed circles). Cultured hypothalamic neurons released GnRH in a pulsatile manner for up to 2 months. As the age of the cultured cells increased from 15– 60 days, there was an increase in peak frequency and a decrease in peak amplitude. The interval between GnRH pulses decreased with duration of culture from 35 6 6.1 min to 26.4 6 15.7 min, 13.7 6 4.1 min, and 15.4 6 4.5 min at 15 days, 30 days, 45 days, and 60 days, respectively. The mean peak amplitude increased from 14.4 6 2.3 pg/ml on day 15 to 38.0 6 2.4 pg/ml on day 30, and decreased thereafter to 10.2 6 4.4 pg/ml and 3 6 1.4 pg/ml at day 60 (Fig. 5). Analysis of the dynamic profile of GnRH release in perifused hypothalamic cultures and GT1–7 cells revealed more complex actions of GnRH agonist and antagonist analogs on the secretory pattern (Fig. 6). First, sustained agonist activation of hypothalamic GnRH receptors by continuous exposure to 1 nm [d-Ala6]Ag extended the interpulse period from

1427

FIG. 3. Actions of GnRH agonist and antagonist analogs on GnRH release from static cultures of hypothalamic cells and GT1–7 cells. Continuous exposure to 50 pM [D-Ala6]Ag had no significant action on GnRH release in cultured hypothalamic cells (A) and GT1–7 cells (B). In contrast, addition of 50 pM [D-pGlu]Antag after the first 48 h of culture caused prominent increases in GnRH release in both hypothalamic and GT1–7 cell cultures (A and B). Asterisks indicate significant differences from the lowest level of GnRH in corresponding groups. Data are representative of three similar experiments.

25.0 6 4.9 min to 33.3 6 6.3 min, and increased the average peak height from 17.1 6 2.1 pg/ml to 25.5 6 2.3 pg/ml (P , 0.01, Fig. 6A). More prominent reduction of spikes frequency and increases in spike amplitude were observed during sustained exposure of hypothalamic cells to 10 nm [d-Ala6]Ag. The interpeak interval increased from 22.5 6 10.0 min to 50.0 6 8.5 min during 3 h of treatment, and the mean peak height increased from 4.9 6 1.7 pg/ml to 13.8 6 2.1 pg/ml. Exposure of perifused hypothalamic cells to 100 nm [d-Ala6]Ag caused a further decrease in pulse frequency, and only a single large peak with mean height of 35.5 6 3.2 pg/ml was observed (Fig. 6B). Thus, increases in agonist concentration caused an increase in the interpeak interval and amplitude of GnRH pulses, leading to less frequent but more prominent episodes of GnRH release. As shown in Fig. 6C, the GnRH agonist analog did not cross-react in the GnRH RIA over a wide range of concentrations. In contrast to the modulatory actions of GnRH agonist on the pattern of neuropeptide secretion, treatment of hypothalamic and GT1–7 cells with SB-75 and [d-pGlu]Antag, respectively, was rapidly followed by cessation of the basal mode of episodic GnRH release, and subsequently by a progressive rise in GnRH release. The latter response was consistent with the above observations in static cultures. In perifused hypothalamic cells, the mean GnRH level (Fig. 7A) rose from the basal value of 8.1 6 0.9 pg/ml to 14.8 6 2.0 pg/ml during the 3-h treatment period. The progressive increase in GnRH release during prolonged antagonist treatment was followed by a prominent peak during washout of SB-75 and returned to near-control levels. Similar increases in basal

1428


Endo • 1999 Vol 140 • No 3

FIG. 4. Pulsatile release of GnRH from cultured hypothalamic neurons. A, Data from perifused hypothalamic cells plated on cytodex beads after 2 weeks in culture. Flow rate, 10 ml/h; collection time, 1 min. B, GnRH secretory profile derived from every fifth sample of the data set shown in (A). C, GnRH secretory profile derived as cumulative production over 5-min periods. Open circles, basal GnRH release; closed circles, depolarization with 35 mM KCl. D, Biological activity of GnRH released by hypothalamic cells, determined by passing perifusion medium from hypothalamic cells over cultured pituitary cells. In A–C, GnRH pulses detected by cluster analysis are indicated by asterisks.

GnRH release were observed in six such experiments during treatment with the potent GnRH antagonist (not shown). In perifused GT1–7 neurons, treatment with the [d-pGlu]Antag also increased GnRH release and was followed by a transient peak response (Fig. 7B). As shown in Fig. 7C, neither of the GnRH antagonist analogs cross-reacted in the GnRH RIA over a wide range of concentrations. Discussion

GnRH-producing neurons originate from the olfactory placode (27) and after their migration into the brain (28) are present in relatively low numbers in the hypothalamus (29). The episodic mode of neuropeptide secretion that is essential for the maintenance of reproductive processes in vertebrates is driven by a hypothalamic GnRH pulse generator (30). The secretory activity of the GnRH neuron is influenced by neurotransmitters, catecholamines, opiates, neuropeptides, pituitary hormones, and gonadal steroids (31–34). In vivo, GnRH-producing neurons make synaptic connections with each other and with numerous other neurons, and are believed to form a neuronal network that is responsible for episodic GnRH secretion (35, 36). The complexity of this structural arrangement has been a limiting factor in the investigation of neuronal pathways, endogenous ligands, and drugs that directly influence the GnRH-producing neurons. In the present studies, all neuronal pathways and interconnections within the hypothalamus were disrupted by dispersion and culture of the hypothalamic cell population. Nevertheless, pulsatile GnRH secretion from such cultured cells was reestablished in vitro and resembled the profile of GnRH release from intact hypothalamic explants (8) and that observed in pituitary portal blood vessels (37). Dispersed hypothalamic cells attached to cytodex beads retain the ability to form interconnections and continue to generate a neuropeptide secretory pattern similar to that observed in vivo

(38). The persistence of such a pattern, with changes in pulse frequency and amplitude during 2 months in culture, indicates that neuropeptide release from a reconstituted neuronal network can occur in the absence of inputs from extrahypothalamic neurons and peripheral endocrine glands. Several studies have shown that immortalized GnRH-producing neuronal cell lines (GT1–1 and GT1–7) also exhibit an oscillatory pattern of GnRH release (31, 32, 39). These observations, and the present findings in cultured hypothalamic cells, suggest that rhythmic activity is an intrinsic property of the GnRH neurons. We have previously reported that GnRH receptors are expressed in GT1–7 cells and that their activation by GnRH agonist analogs influences the pattern of episodic GnRH release from perifused cultures. Other studies have shown that GnRH receptors are widely distributed throughout the brain, and that their density and mRNA levels fluctuate during the estrous cycle (40, 41). Considerable overlapping of the brain areas that contain GnRH-producing cells, and those that exhibit expression of GnRH receptor mRNA, has been reported (42). These findings are consistent with the possibility that an ultrashort-autocrine feedback mechanism may contribute to the control of GnRH secretion from the hypothalamus (20). This hypothesis has been supported by our present finding that immunocytochemically identified GnRH neurons in cultured hypothalamic cells also express GnRH-R that are demonstrable by immunostaining with a receptor antiserum. Also, GnRH-R transcripts were present in cultured hypothalamic cells as well as in hypothalamic tissue from adult animals, and their molecular characteristics were similar to those found in the pituitary gland and aT3 cells (43). These receptors bind native GnRH and its potent analogs in a similar manner to the GnRH receptors that are present in GT1 cells and aT3 gonadotrophs (20). The expression of GnRH receptors in hypothalamic GnRH neurons


1429

FIG. 5. Effects of culture duration on the quantity and pattern of basal GnRH release. A, Time-dependent changes in the amount of released GnRH. B, Peak frequency at increasing durations of primary culture. A similar pattern of basal GnRH release was observed in three individual experiments.

provides a physiological basis for a number of earlier published observations on feedback actions of GnRH on its secretion in vivo. Thus, indirect evidence for the operation of an ultrashort feedback mechanism in the control of GnRH secretion has been obtained by the analyzing the effects of intraventricular injection of GnRH on LH release (44, 45). More direct evidence for an autocrine action of GnRH was obtained from studies with GT1–7 cells, in which rapid increases in [Ca21]i were elicited by treatment with a potent GnRH agonist (20). In addition, activation of GnRH receptors caused transient hyperpolarization that was followed by recovery of electrical activity with increased spike frequency (46). In perifused GT1–7 cells, treatment with GnRH agonists also increased GnRH release, consistent with the rapid elevation in [Ca21]i, and caused a dose-dependent change in pulse amplitude and frequency. Increasing agonist concentrations caused an increase in peak amplitude, with prolongation of the interpeak interval reduced basal GnRH oscillations (20). In the present study, the receptor-mediated actions of GnRH agonists on the neuronal network of cultured hypothalamic cells likewise caused a decrease in frequency and an increase in amplitude of the GnRH pulses. Thus, autocrine feedback leads to switching of the basal pulsatile pattern of release to one characterized by less frequent but more prominent episodes of GnRH secretion. This observation demonstrates that the modulatory actions of GnRH receptor activation on neurosecretion from immortalized GnRH neurons are also evident in the neural network formed in cultured hypothalamic cells. It is of interest that increased release of GnRH from the stalk-median eminence

FIG. 6. Dose-dependent actions of agonist treatment on GnRH release from perifused hypothalamic cells. An initial 2-h period of basal GnRH release (open circles) was followed by exposure to 1 nM (A) and 100 nM (B) [D-Ala6]Ag for 3 h (closed circles). C, Absence of crossreactivity of the [D-Ala6]Ag in the GnRH RIA. Data are representative of seven similar experiments.

region of ovariectomized rhesus monkeys has been observed during prolonged agonist infusion (47). However, in ovariectomized rats the GnRH concentration in hypophyseal portal plasma was reduced by GnRH agonist treatment (45). These contradictory results could reflect differences in the experimental models and time of sampling, since both stimulatory and inhibitory effects were observed during longterm sample collection. In contrast to the agonist-induced changes in the pattern of GnRH release from hypothalamic neurons and GT1 cells, treatment with specific receptor antagonists such as SB-75 (48) and [d-pGlu]Antag caused substantial increases in GnRH release from both cell types. In static cultures, this was manifested over several days in the presence of low antagonist concentrations. In perifused hypothalamic neurons and GT1 cells, blockade of GnRH receptors abolished the basal episodic release of GnRH and initiated a prolonged phase of nonpulsatile neuropeptide secretion. These findings indicate that autocrine activation of GnRH receptors in GnRH neurons is an important component of pulsatile GnRH secretion. Furthermore, the progressive rise in GnRH release during sustained antagonist blockade suggests that autocrine acti-

1430


Endo • 1999 Vol 140 • No 3

ultrashort-loop autocrine feedback mechanisms. The existence of such autoregulatory actions of GnRH is relevant to the maintenance and control of the episodic mode of gonadotropin secretion. The operation of such an oscillator in vivo is influenced by hormonal modulation from peripheral endocrine glands, neuropeptides, and neurotransmitters, by activation of specific plasma-membrane receptors and channels. GnRH neurons are also modulated through synaptic and gap connections within the neuronal network that connects hypothalamic and extrahypothalamic regions. Within such a complex regulatory system, the intrinsic oscillatory capacity of GnRH-producing neurons provides the basal mode of pulsatile GnRH release, and permits the generation of the midcycle LH surge that triggers ovulation. References

FIG. 7. Effects of GnRH receptor blockade on GnRH release from perifused hypothalamic cells. A, After 2 h of basal GnRH release (open circles), the GnRH receptor antagonist, SB-75 (10 nM), was applied for the subsequent 3-h recording period (closed circles). B, Actions of [D-pGlu]Antag (50 pM, closed circles) on GnRH release from perifused GT1–7 Cells. C, Lack of cross-reactivity of SB-75 and [D-pGlu]Antag in the GnRH RIA. Data are representative of six similar experiments.

vation of neuronal GnRH receptors by low GnRH concentrations periodically inhibits the constitutive release of GnRH from the GnRH neuronal network. Such constitutive secretion is probably dependent on the intrinsic excitability of the GnRH neuron, as indicated by the spontaneous firing of such neurons (14, 15, 46) and the ability of tetrodotoxin to inhibit basal GnRH secretion (39). The extent to which such autocrine inhibitory actions of GnRH are related to the activation of inhibitory G proteins (49, 50), and the suppression of second messenger signaling via the calcium and/or cAMP pathways to regulate exocytosis, has yet to be determined. It is noteworthy that increases in the size and frequency of GnRH pulses have been observed in ovariectomized ewes during GnRH antagonist administration (51). The present findings, and observations in other experimental models, indicate that normal GnRH-producing neurons coexpress GnRH and GnRH receptors and exhibit spontaneous electrical activity that controls basal GnRH release. These characteristics are appropriate for the operation of an oscillator that is controlled by both positive and negative

1. Martini L, Fraschini F, Motta M 1968 Comments on long and short feedback loops. In: Michael PR (ed) Endocrinology and Human Behavior. Oxford University Press, London, pp 175–185 2. Motta M, Fraschini F, Martini L 1969 Short feedback mechanisms in the control of anterior pituitary function. In: Ganong WF, Martini L (eds) Frontiers in Neuroendocrinology. Oxford University Press, New York, pp 221–253 3. Sarkar DK 1987 In vivo secretion of LHRH in ovariectomized rats is regulated by a possible autofeedback mechanism. Neuroendocrinology 45:510 –513 4. Chen HT, Geneau J, Meites J 1977 Effects of castration, steroid replacement, and hypophysectomy on hypothalamic LHRH and serum LH. Proc Soc Exp Biol Med 156:127–131 5. Hyyppa M, Motta M, Martini L 1971 ’Ultrashort’ feedback control of folliclestimulating hormone-releasing factor secretion. Neuroendocrinology 7:227–235 6. Kim K, Ramirez VD 1986 In vitro LHRH release from superfused hypothalamus as a function of the rat estrous cycle: effect of progesterone. Neuroendocrinology 42:392–398 7. Melrose PA 1987 In vitro evidence for short-loop gonadotropin feedback on gonadotropin-releasing hormone neurons harvested from adult male rats. Endocrinology 121:200 –204 8. Bourguignon JP, Gerard A, Mathieu J, Mathieu A, Franchimont P 1990 Maturation of the hypothalamic control of pulsatile gonadotropin-releasing hormone secretion at onset of puberty. I. Increased activation of N-methyl-daspartate receptors. Endocrinology 127:873– 881 9. Wetsel WC, Valenca MM, Merchenthaler I, Liposits Z, Lopez FJ, Weiner RI, Mellon PL, Negro-Vilar A 1992 Intrinsic pulsatile secretory activity of immortalized luteinizing hormone-releasing hormone-secreting neurons. Proc Natl Acad Sci USA 89:4149 – 4153 10. Martinez de la Escalera G, Choi AL, Weiner RI 1992 Generation and synchronization of gonadotropin-releasing hormone (GnRH) pulses: intrinsic properties of the GT1–1 GnRH neuronal cell line. Proc Natl Acad Sci USA 89:1852–1855 11. Krsmanovic LZ, Stojilkovic SS, Merelli F, Dufour SM, Virmani MA, Catt KJ 1992 Calcium signaling and episodic secretion of gonadotropin-releasing hormone in hypothalamic neurons. Proc Natl Acad Sci USA 89:8462– 8466 12. Kelly MJ, Condon TP, Levine JE, Ronnekleiv OK 1985 Combined electrophysiological, immunocytochemical and peptide release measurements in the hypothalamic slice. Brain Res 345:264 –270 13. Bosma MM 1993 Ion channel properties and episodic activity in isolated immortalized gonadotropin-releasing hormone (GnRH) neurons. J Membr Biol 136:85–96 14. Van Goor F, Krsmanovic LZ, Catt KJ, Stojilkovic SS, Control of action potential-driven calcium influx in GT1 neurons by the activation status of sodium and calcium channels. Mol Endocrinol, in press 15. Kusano K, Fueshko S, Gainer H, Wray S 1995 Electrical and synaptic properties of embryonic luteinizing hormone-releasing hormone neurons in explant cultures. Proc Natl Acad Sci USA 92:3918 –3922 16. Charles AC, Kodali SK, Tyndale RF 1996 Intercellular calcium waves in neurons. Mol Cell Neurosci 7:337–353 17. Charles AC, Hales TG 1995 Mechanisms of spontaneous calcium oscillations and action potentials in immortalized hypothalamic (GT1–7) neurons. J Neurophysiol 73:56 – 64 18. Matesic DF, Germak JA, Dupont E, Madhukar BV 1993 Immortalized hypothalamic luteinizing hormone-releasing hormone neurons express a connexin 26-like protein and display functional gap junction coupling assayed by fluorescence recovery after photobleaching. Neuroendocrinology 58:485– 492 19. Wilson RC, Kesner JS, Kaufman JM, Uemura T, Akema T, Knobil E 1984 Central electrophysiologic correlates of pulsatile luteinizing hormone secretion in the rhesus monkey. Neuroendocrinology 39:256 –260

AUTOCRINE REGULATION OF GnRH SECRETION 20. Krsmanovic LZ, Stojilkovic SS, Mertz LM, Tomic M, Catt KJ 1993 Expression of gonadotropin-releasing hormone receptors and autocrine regulation of neuropeptide release in immortalized hypothalamic neurons. Proc Natl Acad Sci USA 90:3908 –3912 21. Ramirez VD, Pickle RL, Lin WW 1991 In vivo models for the study of gonadotropin and LHRH secretion. J Steroid Biochem Mol Biol 40:143–154 22. Mellon PL, Windle JJ, Goldsmith PC, Padula CA, Roberts JL, Weiner RI 1990 Immortalization of hypothalamic GnRH neurons by genetically targeted tumorigenesis. Neuron 5:1–10 23. Chomczynski P, Sacchi N 1987 Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162:156 –159 24. Sambrook J, Fritsch EF, Maniatis T Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 25. Reinhart J, Mertz LM, Catt KJ 1992 Molecular cloning and expression of cDNA encoding the murine gonadotropin-releasing hormone receptor. J Biol Chem 267:21281–21284 26. Urban RJ, Johnson ML, Veldhuis JD 1989 In vivo biological validation and biophysical modeling of the sensitivity and positive accuracy of endocrine peak detection. I. The LH pulse signal. Endocrinology 124:2541–2547 27. Schwanzel-Fukuda M, Pfaff DW 1989 Origin of luteinizing hormone-releasing hormone neurons. Nature 338:161–164 28. Jennes L, Conn PM 1994 Gonadotropin-releasing hormone and its receptors in the rat brain. Front Neuroendocrinol 15:51–77 29. Kim K, Ramirez VD 1985 Dibutyryl cyclic adenosine monophosphate stimulates in vitro luteinizing hormone-releasing hormone release only from median eminence derived from ovariectomized, estradiol-primed rats. Brain Res 342:154 –157 30. O’Byrne KT, Thalabard JC, Grosser PM, Wilson RC, Williams CL, Chen MD, Ladendorf D, Hotchkiss J, Knobil E 1991 Radiotelemetric monitoring of hypothalamic gonadotropin-releasing hormone pulse generator activity throughout the menstrual cycle of the rhesus monkey. Endocrinology 129:1207–1214 31. Stojilkovic SS, Krsmanovic LZ, Spergel DJ, Catt KJ 1994 Gonadotropinreleasing hormone neurons intrinsic pulsatility and receptor-mediated regulation. Trends Endocrinol Metab 5:201–209 32. Krsmanovic LZ, Stojilkovic SS, Catt KJ 1996 Pulsatile gonadotropin-releasing hormone release and its regulation. Trends Endocrinol Metab 7:56 –59 33. Urban JH, Das I, Levine JE 1996 Steroid modulation of neuropeptide Yinduced luteinizing hormone releasing hormone release from median eminence fragments from male rats. Neuroendocrinology 63:112–119 34. Lin WW, Ramirez VD 1992 Influences of castration and testosterone on spring to summer changes in release of luteinizing-hormone-releasing hormone in rabbits. J Reprod Fertil 96:1–11 35. Jennes L, Stumpf WE, Sheedy ME 1985 Ultrastructural characterization of gonadotropin-releasing hormone (GnRH)-producing neurons. J Comp Neurol 232:534 –547 36. Leranth C, Segura LM, Palkovits M, MacLusky NJ, Shanabrough M, Naftolin F 1985 The LH-RH-containing neuronal network in the preoptic area of

37. 38. 39. 40. 41.

42. 43. 44. 45. 46.

47. 48. 49. 50. 51.

1431

the rat: demonstration of LH-RH-containing nerve terminals in synaptic contact with LH-RH neurons. Brain Res 345:332–336 Levine JE, Duffy MT 1988 Simultaneous measurement of luteinizing hormone (LH)-releasing hormone, LH, and follicle-stimulating hormone release in intact and short-term castrate rats. Endocrinology 122:2211–2221 Lin WW, Ramirez VD 1991 Seasonal changes in the in-vivo activity of the luteinizing hormone-releasing hormone (LHRH) neural apparatus of male rabbits monitored with push-pull cannulae. J Reprod Fertil 91:531–542 Weiner RI, Wetsel W, Goldsmith P, Martinez de la Escalera G, Windle J, Padula C, Choi A, Negro-Vilar A, Mellon P 1992 Gonadotropin-releasing hormone neuronal cell lines. Front Neuroendocrinol 13:95–119 Jennes L, Eyigor O, Janovick JA, Conn PM 1997 Brain gonadotropin releasing hormone receptors: localization and regulation. Recent Prog Horm Res 52:475– 490 Seong JY, Kang SS, Kam K, Han YG, Kwon HB, Ryu K, Kim K 1998 Differential regulation of gonadotropin-releasing hormone (GnRH) receptor expression in the posterior mediobasal hypothalamus by steroid hormones: implication of GnRH neuronal activity. Brain Res Mol Brain Res 53:226 –235 Jennes L, McShane T, Brame B, Centers A 1996 Dynamic changes in gonadotropin releasing hormone receptor mRNA content in the mediobasal hypothalamus during the rat estrous cycle. J Neuroendocrinol 8:275–281 Stojilkovic SS, Reinhart J, Catt KJ 1994 Gonadotropin-releasing hormone receptors: structure and signaling transduction pathways. Endocr Rev 15:462– 499 Bedran de Castro JC, Khorram O, McCann SM 1985 Possible negative ultrashort loop feedback of luteinizing hormone releasing hormone (LHRH) in the ovariectomized rat. Proc Soc Exp Biol Med 179:132–135 Valenca MM, Johnston CA, Ching M, Negro-Vilar A 1987 Evidence for a negative ultrashort loop feedback mechanism operating on the luteinizing hormone-releasing hormone neuronal system. Endocrinology 121:2256 –2259 Zheng L, Krsmanovic LZ, Vergara LA, Catt KJ, Stojilkovic SS 1997 Dependence of intracellular signaling and neurosecretion on phospholipase D activation in immortalized gonadotropin-releasing hormone neurons. Proc Natl Acad Sci USA 94:1573–1578 Pu S, Terasawa E 1994 Effects of an LHRH agonist analog infusion on the pulsatile release of LHRH and NPY in ovariectomized rhesus monkeys. The Endocrine Society 521 (Abstract) Pinski J, Yano T, Groot K, Milovanovic S, Schally AV 1992 Comparison of biological effects of a sustained delivery system and nonencapsulated LH-RH antagonist SB-75 in rats. Peptides 13:905–911 Mores N, Krsmanovic LZ, Catt KJ 1996 Activation of LH receptors expressed in GnRH neurons stimulates cyclic AMP production and inhibits pulsatile neuropeptide release. Endocrinology 137:5731–5734 Krsmanovic LZ, Mores N, Navarro CE, Saeed S Abdul, Arora KK, Catt KJ 1998 Muscarinic regulation of intracellular signaling and neurosecretion in GnRH neurons. Endocrinology 139:4037– 4043 Padmanabhan V, Evans NP, Dahl GE, McFadden KL, Mauger DT, Karsch FJ 1995 Evidence for short or ultrashort loop negative feedback of gonadotropinreleasing hormone secretion. Neuroendocrinology 62:248 –258