Effects of Changing Gonadotropin-Releasing Hormone Pulse Frequency and Estrogen Treatment on Levels of Estradiol Receptor- and Induction of
http://www.100md.com
内分泌学杂志 2005年第3期
Prince Henry’s Institute of Medical Research, Clayton, Victoria 3168, Australia
Address all correspondence and requests for reprints to: Professor Iain J. Clarke, Prince Henry’s Institute of Medical Research, P.O. Box 5152, Clayton, Victoria 3168, Australia. E-mail: iain.clarke@phimr.monash.edu.au.
Abstract
Estrogen receptor- (ER) levels in gonadotropes are increased during the follicular phase of the ovine estrous cycle, a time of increased frequency of pulsatile secretion of GnRH and elevated plasma estrogen levels. In the present study, our first aim was to determine which of these factors causes the rise in the number of gonadotropes with ER. Ovariectomized hypothalamo-pituitary disconnected ewes (n = 4–6) received the following treatments: 1) no treatment, 2) injection (im) of 50 μg estradiol benzoate (EB), 3) pulses (300 ng iv) of GnRH every 3 h, 4) GnRH treatment as in group 3 and EB treatment as in group 2, 5) increased frequency of GnRH pulses commencing 20 h before termination, and 6) GnRH treatment as in group 5 with EB treatment. These treatments had predictable effects on plasma LH levels. The number of gonadotropes in which ER was present (by immunohistochemistry) was increased by either GnRH treatment or EB injection, but combined treatment had the greatest effect. Immunohistochemistry was also performed to detect phosphorylated cAMP response element binding protein (pCREB) and Fos protein in gonadotropes. The number of gonadotropes with Fos and with pCREB was increased only in group 6. We conclude that either estrogen or GnRH can up-regulate ER in pituitary gonadotropes. On the other hand, during the period of positive feedback action of estrogen, the appearance of pCREB and Fos in gonadotropes requires the combined action of estrogen and increased frequency of GnRH input. This suggests convergence of signaling for GnRH and estrogen.
Introduction
THE DECLINE IN plasma progesterone levels at the end of the luteal phase of the estrous cycle removes negative feedback on GnRH, allowing for increased frequency of pulsatile discharges of this neurohormone from the hypothalamus (1, 2). This, in turn, increases the frequency of pulsatile discharges of LH from the pituitary gonadotropes leading to growth of ovarian follicles and increased estrogen secretion (3, 4). The increase in circulating levels of estrogen induces a positive feedback effect on the secretion of GnRH and LH, which is a time-delayed mechanism (4). This involves actions of estrogen on cells in the brain as well as action on the pituitary gonadotropes, involving electrophysiological properties and mobilization of secretory granules (3, 5, 6, 7, 8). The specific molecular actions of estrogen that produce the GnRH/LH surge are not well understood, although this is the primary endocrine signal for ovulation. This estrogenic signal that initiates positive feedback occurs in the follicular phase of the normal estrous cycle (4).
Although there is a wealth of literature regarding the subcellular mechanisms for GnRH action in gonadotropes, there is relatively little information on the means by which estrogen acts on these cells. Estrogen regulates the expression of GnRH receptors in vivo (9, 10, 11) and in vitro (12), with an increase occurring during the follicular phase of the estrous cycle in the ewe (13, 14, 15, 16). Nevertheless, responsiveness of the gonadotropes to GnRH does not increase until close to the time of onset of the LH surge (1), suggesting that this increased GnRH receptor expression may be permissive of enhanced secretion at the time of the LH surge, but it is not the cause of a profound increase in responsiveness at the onset of the surge. LH-secreting granules migrate to the periphery of gonadotropes in preparation for the cyclic LH surge (17), which presumably reflects subcellular mechanisms activated by estrogen. This polarization of the granules also occurs in ovariectomized (OVX) ewes given a single im injection of estradiol benzoate (EB) (7).
A series of studies on rat pituitaries show that there is up-regulation of second messenger systems in the proestrous stage of the estrous cycle, suggesting that these factors may be related to the mechanism of positive feedback (18, 19, 20, 21, 22). Because these studies were done with hemipituitaries, however, and various cell types in the rat pituitary express ER (23), it is difficult to ascribe the activities directly to the gonadotrope. One study in particular provides convincing evidence for the action of estrogen to up-regulate protein kinase C in rat pituitaries (18). These data suggest actions of estrogen that may either be genomic or nongenomic, and recent data (24) further indicate rapid effects via phosphorylation of MAPK and phosphorylation of cAMP response element binding protein (CREB). Short- and long-term effects of estrogen on CREB and related proteins, as well as CRE-DNA binding, are seen in various brain regions (25); the former could be caused by rapid nongenomic effects of estrogen (26, 27, 28). Measurement of phosphorylation of CREB is regarded as an index of cellular activation (24, 25). Fos is another marker of cellular response, and GnRH has been shown to increase the percentage of immuno-identified gonadotropes that produce Fos protein (29).
Whereas estrogen up-regulates GnRH receptor levels (see above), we showed that the number of gonadotropes that express estrogen receptor- (ER) is also increased in the follicular phase of the ovine estrous cycle (30). This could be a means of enhancing sensitivity to estrogen at this time. Early work in rat cells (31) suggested that GnRH could increase nuclear ER in pituitary cells. We hypothesized that the increased expression of ER in gonadotropes in the follicular phase of the estrous cycle might be caused by increased frequency of GnRH input, increasing plasma levels of estrogen, or both. To investigate these possibilities, we employed our well characterized model of the OVX hypothalamo-pituitary disconnected (HPD) ewe, which allows the study of the pituitary gland without brain inputs (32, 33). Programmed replacement with pulsatile GnRH and estrogen allows the study of effects on the gonadotrope in vivo that can be correlated with subsequent postmortem analysis (34). We also examined the effects of either increased GnRH pulse frequency and/or estrogen treatment on the appearance of phosphorylated CREB (pCREB) and Fos in the gonadotropes. The treatment paradigm was such that the measurements were taken at the time when positive feedback actions of estrogen are operative (20 h after an im injection of EB) (33).
Materials and Methods
Ethics statement
All experimental procedures were conducted under a protocol approved by the Victorian Institute of Animal Science Animal Ethics Committee.
Sheep
Adult Corriedale female sheep used in this study were kept at the Prince Henry’s Institute facility at Werribee, Melbourne. The animals were bilaterally OVX at least 1 month before the experiment, and the HPD operation was performed as previously described (32). The latter procedure removes all neural inputs to the median eminence, but gonadotropin secretion can be restored by iv administration of GnRH in a pulsatile mode. Immediately after HPD, one external jugular vein was cannulated with a 12-gauge Teflon Dwellcath (Tuta Laboratories Australia Pty. Ltd., Bedford Park, South Australia, Australia) that was connected to a manometer line (Portex Ltd., Hythe, UK) for infusion of GnRH or vehicle (see below). These infusion lines were connected to a Braun Unita 1 infusion pump that could be programmed to deliver 2.25-ml pulses over 6 min as previously described (34). All animals received vaginal progesterone delivery devices (controlled intravaginal drug release) (Lyppard, Cheltenham, Victoria, Australia) for 5 d with GnRH or saline vehicle replacement as detailed below. These were removed at 1600 h on the fifth day and EBL or oil vehicle (1 ml peanut oil) was administered at 1600 h on the sixth day. The animals were killed for recovery of pituitary glands at 1200 h on the seventh day (20 h after EB or oil vehicle).
The animals were assigned to experimental groups as follows (summarized in Fig 1): 1) saline vehicle and oil vehicle (n = 5); 2) saline vehicle and injection (im) of 50 μg EB (Intervet, Baulkham Hills, New South Wales, Australia) (n = 5); 3) pulsatile replacement of GnRH by iv infusion of 300 ng every 3 h with oil vehicle (n = 6); 4) same GnRH regimen as group 3 with EB treatment (n = 4); 5) pulsatile replacement of GnRH by iv infusion of 300 ng every 3 h, switched to hourly pulses of 100 ng between 1600 h on d 6 until 0800 h on d 7, and then switched to 50-ng pulses every 0.5 h between 0800 h on d 7 until 1200 h on d 7 (time of termination), oil vehicle treatment (n = 4); and 6) same GnRH regimen as group 5 with EB treatment (n = 4).
FIG. 1. Experimental design.
The dose of each GnRH pulse was varied in accordance with the interval of delivery, such that the animals received the same amount of GnRH per unit time. In this way, groups 3 and 4 received GnRH replacement at a low frequency, characteristic of that seen during the luteal phase of the estrous cycle (1, 35). Groups 5 and 6 received GnRH replacement where the pulsatile input increased in frequency, as seen in the follicular phase of the estrous cycle. Previous work has shown that the EB treatment of GnRH-replaced OVX-HPD ewes causes a biphasic response in terms of LH secretion, with a time frame similar to that seen in the OVX, hypothalamo-pituitary intact ewe (36, 37). A surge in LH secretion is seen in OVX-HPD ewes between 16 and 20 h after EB treatment, but this reaches only the magnitude of that seen in the hypothalamo-pituitary intact animal if the GnRH pulse frequency is increased to 10-min intervals or if a large bolus of GnRH is delivered at 16 h (36, 37).
Blood samples (5 ml) were taken as follows. On d 6, all animals were sampled over three GnRH pulses delivered at 0900, 1200, and 1500 h; samples were taken 5 min before the delivered pulse and at 7.5, 15, and 30 min after the pulses. For animals in groups 5 and 6, samples were collected over three pulses on d 6 and then over hourly pulses between 0500 and 0700 h and then over half-hourly pulses between 0800 and 1200 h. For sampling around half-hourly pulses on d 7, samples were taken 7.5, 15, and 25 min after the pulses. For groups 1–4, samples were collected at the same times as indicated for groups 5 and 6, even though GnRH or saline pulses were delivered at 3-hourly intervals in these animals; this enabled quantification of changes in total LH released over the last 3 h of the experiment.
The animals were overdosed with sodium pentobarbital (0.5 ml/kg body weight, iv Lethobarb, May and Baker Pty. Ltd., Melbourne, Australia), and the pituitaries were carefully removed from the sella turcica, cut into 3-mm sagittal slices, and placed in a 4% paraformaldehyde in 0.1 M phosphate buffer (PB) (pH 7.4) for 2 d at 4 C. The tissues were then placed in 30% sucrose in 0.1 M PB solution for 5 d at 4 C and then frozen on powdered dry ice and stored at –20 C.
Immunohistochemistry
For all immunohistochemical procedures, antibody incubations were carried out at room temperature unless otherwise specified and were followed by three 10-min washes in 0.05 M PBS. Pituitary sections (7 μm) were cut on a cryostat (Leica CM 1850) at –22 C, thaw-mounted onto Superfrost slides, and stored at –20 C until used for immunohistochemistry. A modified version of a recently published antigen retrieval method was used to detect ER in ovine pituitary tissue (30). This was preceded by incubation with 3% H2O2 in methanol (30 min) to block endogenous peroxidase activity. Slides were then microwaved (four times for 5 min at 700 W) in 0.01 M citrate buffer (pH 6.0) and left to stand in the same buffer for another 20 min. For immunocytochemistry, the sections were incubated in blocking solution (10% normal donkey serum, 0.3% Triton X-100 in 0.1 M PB) for 30 min and then with ER antibody (monoclonal ID5 raised against the N-terminal region of ER, Dako, Carpinteria, CA) at a 1:8 dilution in blocking solution for 72 h at 4 C. The sections were then incubated (1 h) in biotinylated donkey antimouse IgG (Vector Laboratories, Burlingame CA) at a 1:400 dilution and then with Streptavidin horseradish peroxidase (STA-HRP) complex (Amersham, Piscataway, NJ) at a 1:500 dilution for 1 h. After washes in buffer, the sections were preincubated in NiDAB (0.5 mg/ml, 0.02% NH4NiSO4 and 0.026% CoCl2 in 0.1 M PB) for 10 min. This solution was removed and 0.15% H2O2 quickly added for approximately 4 min. The color reaction was monitored and halted by removing the NiDAB solution and washing in 0.1 M PB. Double labeling was then carried out to identify LH cells that also expressed ER. Expression of LH was detected using an anti-LH polyclonal antibody (30) at 1/5000 and visualized with diaminobenzidine (DAB) to give a brown color. Briefly, after immunohistochemistry for ER, sections were washed in 0.05 M PBS and incubated for 30 min in blocking solution containing normal goat serum, then with primary antibody for 48 h at 4 C, followed by incubation (1 h) in secondary antibody (biotinylated goat antirabbit, Vector) at 1/200 dilution. Sections were then incubated with STA-HRP complex for 1 h followed by DAB solution. The color reaction was as described above but without the addition of NH4NiSO4 and CoCl2.
To identify activated LH cells, a separate series of Fos/LH double-labeling and pCREB/LH double-labeling immunohistochemistry was carried out on two pituitary sections per animal per group. Polyclonal rabbit antibodies to Fos protein (AB5, Oncogene Research, Cambridge, MA; 1/1000 dilution) and pCREB (Upstate Biotechnology, Placid, NY; 1/500 dilution) were used as markers for early activation; both were visualized with NiDAB. Sections were washed in 0.05 M PBS followed by pretreatment with 3% H2O2 in methanol, washed in PBS, and then processed for antigen retrieval as described above. After washes in PBS, the sections were incubated with blocking solution containing normal goat serum for 30 min and then in primary antibody for 72 h at 4 C. This was followed by incubation (1 h) in biotinylated goat antirabbit IgG at 1/200 dilution and then in STA-HRP complex at 1/500 dilution for 1 h. After washes in PBS, sections were then incubated with NiDAB solution and color development carried out as above. Double labeling was achieved by incubation with polyclonal rabbit anti-LH antibody visualization with DAB as described above. The antibodies used for labeling Fos (38), pCREB (39), and ER (40) have been characterized previously.
Analysis
Five regions (one in each of the four quadrants and one region in the center) of each coronal anterior pituitary section were examined under x60 magnification. At this magnification, the presence or absence of nuclear and cytoplasmic staining was unambiguous. In each region, 100 cells were counted to give a total of 500 cells per section. Single LH-positive cells and double-labeled LH/ER, LH/Fos, and LH/pCREB cells were scored for each series of sections. Mean percent colocalization was then calculated for each experimental group.
LH RIA
Plasma LH levels were measured using a RIA as previously described (38). In six assays, sensitivity was 0.1 ng/ml, within-assay coefficient of variation was 3.5–9.8% between 3.3 and 6.1 ng/ml, and between-assay coefficient of variation was 12% at 12.7 ng/ml, 10% at 5.0 ng/ml, 7.3% at 6.6 ng/ml, and 6% at 23.3 ng/ml.
Statistical analysis
To determine statistically significant (P < 0.05) differences between means, area under curves (AUCs) of plasma LH concentration vs. time, and for the percentages of gonadotropes that colocalized either ER or Fos or pCREB, two-way ANOVA was performed. Post hoc tests for differences between groups employed the method of least-significant differences. For analysis of percentage data, the arcsin transformation was used as recommended by Sokal and Rohlf (41). In the case where pretreatment values were compared with posttreatment values, paired Student’s t tests were used.
Results
Plasma LH levels were undetectable in the sheep of groups 1 and 2. Analysis of effects of GnRH with either EB or oil treatment was therefore confined to groups 3–6.
Typical examples of the pattern of response to the different GnRH/EB treatments are seen in Fig. 2. LH pulse amplitudes in the animals given GnRH pulses every 3 h plus oil injection (group 3) decreased (not significant) slightly (14%) over the course of the experiment (Table 1), but the LH pulse amplitudes were reduced by 23% (P < 0.01) and broadened (Fig. 2) in animals given the same GnRH treatment with EB injection, (Table 1). When GnRH pulse frequency was increased (with oil treatment; group 5), the LH pulse amplitudes were reduced across the experiment with progressive increase in GnRH/LH pulse frequency (Table 1). In those animals that received EB with increased pulse frequency (group 6), the LH pulse amplitudes were not significantly reduced on hourly pulse frequency, because one animal showed a marked increase in LH pulse amplitude. When group 6 animals were given half-hourly pulses of GnRH, the LH pulse amplitudes were significantly (P < 0.01) reduced (Table 1).
FIG. 2. Examples of plasma LH levels in individual OVX-HPD ewes receiving GnRH on a fixed 3-hourly pulse frequency or increased pulse frequency, with or without EB treatment. Black bars indicate the timing of GnRH pulses, which were either 300 ng (3-hourly) before EB/oil treatment or 50 ng (half-hourly) during the last 3 h of the animal treatments. Examples are given for animals from groups 3 (upper left), 4 (upper right), 5 (lower left), and 6 (lower right).
TABLE 1. Effect of changing GnRH pulse frequency and effect of EB/oil on LH pulse amplitude in OVX-HPD ewes
AUCs of plasma LH concentration vs. time were analyzed for data obtained between 0900 and 1200 h on d 7, because all animals were sampled at the same clock times. As seen in Fig. 2, the LH pulse shapes were different in animals from groups 1–4, and mean AUC data for groups 4–6 are presented in Table 2. Plasma LH was undetectable in groups 1 and 2, so the analysis excluded these. The data show that EB treatment with increased GnRH pulse frequency (group 6) caused greater release of LH than all other treatments. In the partitioning of effects, the increased GnRH pulse frequency produced significantly (P < 0.01) higher values than the fixed-pulse regimen, and EB treatment produced greater values than oil treatment (P < 0.001). The interaction between the two was significant (P < 0.002), indicating that the combination of increased GnRH pulse frequency with EB treatment gave the greatest AUC values.
TABLE 2. Area under plasma LH concentration vs. time curves between 0900 and 1200 h on d 7 (when groups 5 and 6 received half-hourly GnRH pulses)
Figure 3 shows examples of gonadotrope immunostaining with double labeling for ER, Fos, and pCREB. The percentage of gonadotropes that colocalized ER is shown in Table 3. Both GnRH treatment and EB treatment increased the percentage of gonadotropes that colocalized ER (group 1 vs. all other groups). With GnRH treatment alone (groups 2 and 3), the level of colocalization was the same with both pulse paradigms. Three-hourly GnRH pulses plus EB (group 5) had a greater effect than 3-hourly pulses alone (group 2). EB treatment alone (group 4) was more effective in elevating ER in gonadotropes than GnRH given at the 3-hourly pulse frequency but not more effective than increased GnRH pulse frequency alone (group 3). Increased GnRH pulse frequency plus EB (group 6) was more effective than increased GnRH pulse frequency alone (group 3). The statistical analysis showed that there was a statistically significant (P < 0.001) interaction between the GnRH treatment and the EB treatment, which reflects the fact that EB treatment with increased GnRH pulse frequency gave a much bigger response than EB treatment with fixed GnRH pulse frequency.
FIG. 3. Examples of the ER (row 1), Fos (row 2), and pCREB (row 3) immunostaining in gonadotropes (detected with LH immunostaining). Boxed areas in the first and third columns are enlarged in the second and fourth columns. Open arrows show gonadotropes without dual staining, closed arrows show double labeling of gonadotropes with ER (row 1), Fos (row 2), and pCREB. The arrowheads indicate ER and Fos immunoreactive nuclei of pituitary cells that were not immunoreactive for LH.
TABLE 3. Mean (± SEM) percentage of gonadotropes (LH immunopositive) that colocalized ER, Fos, and pCREB as determined by immunostaining
Both the percentage of gonadotropes that colocalized Fos and the percentage that colocalized pCREB were increased only in the group that received increased GnRH pulse frequency and EB (group 6) (Table 3).
Discussion
Using an in vivo model, we have shown that both GnRH and estrogen increase the number of gonadotropes with immunohistochemically identifiable ER. A complex interaction clearly exists between actions of the GnRH receptor and the ER that allows combined regulation of ER, and this may explain the increase in ER in gonadotropes that is seen in the follicular phase of the estrous cycle (30). Furthermore, we have shown that a combination of increased GnRH pulse frequency and estrogen is required to generate Fos and pCREB responses in these cells. Although earlier studies have shown that GnRH stimulates production of immediate early response genes in T3-1 cells and gonadotropes (29, 42), and estrogen induction of Fos has been observed in the pituitary gland (see below), the combined effect on gonadotropes has not been observed previously. An earlier report (43) demonstrated that in vitro treatment of perifused pituitary cell aggregates with GnRH pulses increased the steady-state level of ER mRNA with no increase in LH? mRNA. These data suggested that there may be convergence of signaling of GnRH and nuclear ER in pituitary gonadotropes, and the present data show that this is indeed the case.
The GnRH and estrogen treatments employed in our study caused predictable changes in the plasma levels of LH, highlighting the utility of the model. In particular, an increase in GnRH pulse frequency reduced LH pulse amplitude, as reported previously (44), and EB altered the shape of LH pulses at the time of predicted positive feedback, as shown previously (33). AUCs were the same for 3-hourly and hourly pulses, as predicted from earlier studies, showing the exactitude of frequency and amplitude relationships (45). The present study confirms that, in the OVX ewe without steroid treatment, the inverse relationship between GnRH pulse frequency and pulse amplitude is rigid and is a reflection of the releasable pool of LH within the cell (45). Although estrogen increased AUC units slightly when animals were given EB on a fixed-pulse frequency, most likely because of altered LH pulse shape, this was not statistically significant. The combination of increased pulse frequency and EB treatment elevated total secretion of LH, although surge-like secretion was not seen. This is because the frequency of GnRH input did not reach that seen during the EB-induced surge in the OVX hypothalamo-pituitary intact ewe (46). Higher-frequency GnRH input in combination with EB treatment does cause a full LH surge in this model (36, 37). The combination of increased GnRH pulse frequency and EB did significantly increase AUC, indicating that the animals were sampled at a time when positive feedback mechanisms were operative. This confirms the effect of estrogen to sensitize cells to respond to GnRH (33). These treatments allow dissection of subcellular events within the gonadotropes that relate to the positive feedback phenomenon.
We employed im injection of EB in oil as an effective means of eliciting the positive feedback phenomenon as previously shown in OVX ewes and in OVX-HPD ewes given pulsatile GnRH replacement (47). The positive feedback effect can also be elicited with iv injection of estradiol to OVX ewes (48), which raises plasma estrogen levels for a period of less than 3 h, and nuclear ER binding (measured by nuclear exchange assay) was elevated for up to 6 h. This indicates that the estrogen signal that initiates the positive feedback effect need only be transient and that a time delay is required before the surge in LH secretion occurs. More recently, others (49) have shown that treatment of OVX ewes with estrogen implants for 8 h will cause an LH surge 17–19 h after commencement of treatment. These authors reinforce the notion that the surge-inducing effect of estrogen involves activation (during which time, estrogen activates mechanisms that will cause a surge with a time delay) and transmission (which is not dependent upon the presence of estrogen). It was also shown in cultured rat pituitary cells by Kamel and Krey (50) that enhancement of LH release by estrogen continues beyond the time that estrogen is withdrawn. It is not known whether liganded ERs are required to be present at the time of the surge, but it is clear that the surge-generating action of estrogen occurs at a time before the surge. Presumably, the delay interval allows a sequence of intracellular signaling that is not well understood. In this paper, we show that activation of at least two transcription factors (Fos and pCREB) require increased pulsatile GnRH input as well as estrogen input. Additional use of this model with biochemical analysis of signaling systems will allow greater resolution of this unique phenomenon.
We previously showed that the number of gonadotropes expressing ER is increased in the follicular phase of the estrous cycle (30), but it was not known whether this was because of increased estrogen levels and/or increased GnRH pulse frequency. The number of gonadotropes that also stain for ER/ER? also increases at proestrus in the rat (51). The present work provides novel in vivo data to suggest that both factors are involved. EB treatment alone caused a significant increase in ER as did 3-hourly pulses of GnRH, but the former was more effective. With increased GnRH pulse frequency in the absence of an EB effect, there was no significant increase in the number of gonadotropes in which immunostaining of ER was seen, suggesting that this manipulation does not lead to an increase in ER. Nevertheless, during the period when a positive feedback effect of EB is operative in our model (33), a combination of EB and increased GnRH pulse frequency increased the percentage of ER-positive gonadotropes to the greatest extent. Thus, the combination is the most effective means of up-regulating ER in this cell type. The follicular phase of the estrous cycle is characterized by increased GnRH/LH pulse frequency (1, 2, 35) as well as increasing levels of estrogen (17). Based on the data of the present study, the increase in gonadotropes that are immunopositive for ER at this time (30) is most likely caused by a combination of these two factors.
Statistical analysis of the effect of GnRH pulse frequency and EB treatment on ER in gonadotropes showed a highly significant interaction between these two parameters, which indicates convergence between estrogen action and GnRH action. Earlier work from various labs and in the ovine model (9) has clearly shown that estrogen up-regulates GnRH receptors, and the present data clearly show that GnRH up-regulates ER. The effect of GnRH on ER protein substantiates the results of Demay et al. (43) (see above). The additive effect could involve action of EB via ER present in these cells even after 7 d of pituitary disconnection without replacement (see group 1 data in Table 3) or action by other means. There are few ER? receptors in the ovine pituitary gland (Scott, C., and I. J. Clarke, unpublished data), so the effect is most likely exclusive to mediation via ER.
GnRH action involves a variety of second messenger systems, including the phospholipases C, D, and A2, protein kinase C, and MAPK pathways (20, 21, 22). Phospholipase hydrolysis of the phosphoinositides also plays a pivotal role, and the release of intracellular calcium is important for the secretory mechanism (20, 52). The cAMP/protein kinase A pathway is operative within gonadotropes, although it has been suggested to be more important for transcriptional regulation than the acute secretory response to GnRH (20). Many of these pathways have also been described in immortalized cell lines. Cell membrane ion channel events are a fundamental component of the secretory response to GnRH (52), and Rab3B appears to be involved (53). Ovine gonadotropes produce a wide range of soluble N-ethyl maleimide-sensitive factor attachment proteins/receptors (8), most likely involved in the exocytosis of gonadotropin-containing vesicles as well. Estrogen can act via ER to directly regulate transcription in gonadotropes (54), and rapid estrogen effects are also possible, involving membrane-associated ER and/or cytoplasmic ER interactions with signaling systems (24, 55, 56). Estrogen sensitization of gonadotropes is a well recognized phenomenon, allowing increased responsiveness to GnRH. This could involve an increase in GnRH receptor number, induced by estrogen but not (at least in the OVX-HPD ewe) by increased GnRH pulse frequency (9, 34). The present model will allow further subcellular dissection of the effects of the combination of estrogen and increased GnRH pulse frequency. A range of other intracellular responses are also seen in the gonadotrope at the time of the estrogen-induced surge in LH secretion (57) that could be the consequence of the genomic action of liganded ER complexes.
Estrogen can act directly on gonadotropes to regulate transcription of the gonadotropin subunits and the GnRH receptor. At least in the case of the ovine FSH?, the promoter region contains sites for the binding of progesterone receptor and activator protein 1 (AP-1) and a site for estrogen-dependent repression (58). The rat LH? gene has an imperfect palindromic sequence in the promoter region that confers estrogen responsiveness (59) allowing direct estrogen action. Similar information is lacking for the ovine LH? gene and the common -subunit, but available evidence shows that estrogen negatively regulates the transcription of all of the gonadotropin subunit genes in the pituitary of the EB-treated OVX ewe over the time course relevant to this study (47). This study also showed that LH? gene expression was not reduced over the period of acute negative and positive feedback after EB treatment of OVX-HPD ewes that were receiving hourly GnRH pulses, presumably because GnRH stimulation was maintained, whereas it declined in the negative-feedback phase in the EB-treated OVX ewe. Steady-state mRNA levels for the GnRH receptor, the -subunit, and LH? (but not FSH?) are increased in the follicular phase of the ovine estrous cycle (13). The human GnRH receptor promoter has a variety of cis-acting regulatory sequences including sites for CRE and AP-1 (60). Estrogen significantly up-regulates GnRH receptor transcription in EB-treated, OVX-HPD ewes receiving constant GnRH input, indicating direct pituitary action (34). Thus, genomic signaling of estrogen can regulate the transcription of various genes within the gonadotrope, and up-regulation of ER by both estrogen and GnRH as shown in this study could facilitate the combined effects of the two hormones. As indicated above, effects on the secretion of LH probably involve a range of signaling systems.
It is also possible that convergence of estrogen and GnRH signaling occurs via membrane-associated ER and/or cytoplasmic ER interactions with signaling systems. In other words, estrogen can act to rapidly phosphorylate CREB and/or MAPK (24), and these could converge on GnRH signaling systems. Recent evidence suggests that ER may associate with calveolin to facilitate membrane binding and signaling via G proteins (61, 62). This can facilitate transactivation of the epidermal growth factor receptor by liganded ER (62). A similar scenario may exist for the interaction of ER and the G protein-coupled GnRH receptor. Such convergence of postreceptor signaling might explain why the increase in GnRH pulse frequency and estrogen act together to activate the gonadotrope or whether effects are solely because of genomic action of liganded ER. More work on this cell type is required before such a mechanism is delineated, and our in vivo HPD model offers a means by which this can be examined in nontransformed gonadotropes.
The appearance of Fos protein is generally regarded as a transient response to a specific activator of cells (63). GnRH produces a Fos response in gonadotropes, and although estrogen has been shown to do the same in a number of different types of cells, there are no published reports detailing the response in gonadotropes. A rapid (within 15 min) response of gonadotropes to GnRH has been seen using Fos as an index (29). An interesting feature of this cell type is that a prolonged Fos response may occur. Thus, Szijan et al. (64) found that a large (250 mg) sc injection of estrogen diundecylenate given to male rats produced elevated c-fos and c-myc mRNA levels in the pituitary gland that were sustained for 16–22 h. Because these authors did not define the cell type and various cell types possess ER in the rat (65), this may or may not relate to gonadotropes. In another study, using 1 μg estradiol given sc to OVX rats, Allen et al. (66) found that c-fos expression was elevated within 2 h and remained elevated for 48 h; this response was localized to lactotropes and folliculostellate cells. Sustained Fos/Fos-related antigen responses are also seen in the dopaminergic cells of the hypothalamus of OVX anestrous ewes chronically treated with estrogen (65). On the other hand, Fos responses to a single im injection of EB cause transient responses in brain cells of the OVX ewe (67). In the present study and with a fixed regimen of GnRH stimulation, the number of gonadotropes that were immunopositive for Fos was similar to untreated gonadotropes. There was no increase in the proportion of gonadotropes with Fos after increased frequency of GnRH input. A transient increase may have occurred earlier in response to the increased pulse frequency, but our design could not test for this. On a fixed GnRH pulse regimen, the number of gonadotropes immunostained for Fos was not increased by EB treatment 20 h earlier, but once again, there could have been an earlier transient response. The most significant feature of the present data is that the combination of increased GnRH pulse frequency and EB treatment significantly increased the number of gonadotropes displaying Fos-immunoreactive nuclei 20 h after the commencement of the combined treatment. This is the period when the positive-feedback effect of estrogen is operative (due to the combined effects of increased GnRH pulse frequency and estrogen), and this is clearly associated with activation of the cells, as indicated by the Fos responses. Because AP-1 binding sites are found in the GnRH receptor promoter (see above), this is at least one gene that could be affected by Fos production. As to how estrogen up-regulates GnRH receptor expression in this experimental model (see above), and during the estrous cycle (13), it is possible that this is via up-regulation of second messenger systems, as indicated above, and Fos.
As with Fos, the level of pCREB was increased only in animals that received increased frequency of GnRH pulses and EB. This could be effected by nongenomic action of estrogen, because rapid effects (within 15 min) of estrogen are seen in the brain for phosphorylation of CREB and MAPK (24). If this involves ER-dependent signaling that is rapid (24), it may involve interaction of the ligand with the receptor in the extranuclear compartment of the cell to phosphorylate signaling molecules directly (55). CREB may act as an integrating point for a range of intracellular signaling systems (68). As indicated above for Fos induction, it is possible that either estrogen or increased GnRH pulse frequency could have caused a transient effect to phosphorylate CREB, but at the time point studied, it was only the combination of the two that produced elevated pCREB production at 20 h. Although estrogen might rapidly phosphorylate CREB through nongenomic mechanisms, chronic estrogen treatment has also been shown to increase pCREB and other components of this system (25), so the increase that we have observed with a combination of increased GnRH pulse frequency could be important for both extranuclear activation of second messenger systems and for genomic effects to regulate transcription. As suggested by Zhou et al. (55), phosphorylation of CREB may allow integration of estrogen signaling and membrane receptor signaling. How this relates to the mechanism of positive feedback remains to be elucidated.
A series of studies by Nett and coworkers (69, 70) show that cAMP production is increased by GnRH treatment in ovine gonadotropes. GnRH stimulates parallel release of LH and cAMP (69). Dibutyryl cAMP was also able to stimulate LH release in their culture system (69). Later work confirmed that dibutyryl cAMP as well as phorbol ester (to stimulate protein kinase C) could stimulate LH and FSH release from ovine pituitary cells in primary culture, but a calcium ionophore selectively stimulated LH release (70). Interestingly, it has been reported that a CRE region is not found on the promoter of the ovine FSH? gene, and it is not known whether such a consensus sequence exists in the ovine LH? or common -subunit gene promoters. As mentioned above, the human GnRH receptor promoter contains a CRE site, so pCREB could be relevant to the transcription of this gene. A range of other second messenger systems (which are activated by estrogen and GnRH; see above) (71) can phosphorylate CREB, leading to effects at the level of transcription.
In summary, we have used a unique in vivo model to elucidate combined effects of increased GnRH pulse frequency and estrogen injection on the levels of ER, Fos, and pCREB in the gonadotropes of the ovine pituitary gland. The data show that the proportion of cells displaying immunoreactive ER is increased by both increased GnRH pulse frequency and estrogen treatment, with interactive effects. This combined action explains the increased level of ER seen during the follicular phase of the estrous cycle (30). In addition, the number of cells displaying Fos and pCREB immunoreactivity is increased by the combined treatment (increased GnRH pulse frequency and estrogen treatment) but not by either alone. It is also this combined treatment that brings about the greatest increase in sensitivity of gonadotropes to GnRH with respect to LH secretion. Again, this suggests an interaction between the signaling systems for GnRH receptors and ERs.
Acknowledgments
We thank Alix Rao, Karen Briscoe, and Bruce Doughton for technical assistance and the National Institutes of Health for assay reagents.
References
Clarke IJ 1995 Evidence that the switch from negative to positive feedback at the level of the pituitary gland is an important timing event for the onset of the preovulatory surge in LH in the ewe. J Endocrinol 145:271–282
Moenter SM, Caraty A, Locatelli A, Karsch FJ 1991 Pattern of gonadotropin-releasing hormone (GnRH) secretion leading up to ovulation in the ewe: existence of a preovulatory GnRH surge. Endocrinology 129:1175–1182
Clarke IJ 1996 The Hypothalamo-pituitary axis. In: Hillier SJ, Kitchener HD, Neilsen JP, eds. Scientific essentials of reproductive medicine. London: WB Saunders; 120–132
Clarke IJ 1995 The preovulatory LH surge: a case of a neuroendocrine switch. Trends Endocrinol Metab 6:241–247
Cowley MA, Chen C, Clarke IJ 1999 Estrogen transiently increases delayed rectifier, voltage-dependent potassium currents in ovine gonadotropes. Neuroendocrinology 69:254–260
Heyward PM, Clarke IJ 1995 A transient effect of estrogen on calcium currents and electrophysiological responses to gonadotropin-releasing hormone in ovine gonadotropes. Neuroendocrinology 62:543–552
Thomas SG, Clarke IJ 1997 The positive feedback action of estrogen mobilizes LH-containing, but not FSH-containing secretory granules in ovine gonadotropes. Endocrinology 138:1347–1350
Thomas SG, Takahashi M, Neill JD, Clarke IJ 1998 Components of the neuronal exocytotic machinery in the anterior pituitary of the ovariectomised ewe and the effects of oestrogen in gonadotropes as studied with confocal microscopy. Neuroendocrinology 67:244–259
Clarke IJ, Cummins JT, Crowder ME, Nett TM 1988 Pituitary receptors for gonadotropin-releasing hormone in relation to changes in pituitary and plasma gonadotropins in ovariectomized hypothalamo/pituitary-disconnected ewes. II. A marked rise in receptor number during the acute feedback effects of estradiol. Biol Reprod 39:349–354
Emons G, Hoffmann HG, Brack C, Ortmann O, Sturm R, Ball P, Knuppen R 1988 Modulation of gonadotropin-releasing hormone receptor concentration in cultured female rat pituitary cells by estradiol treatment. J Steroid Biochem 31:751–756
Menon M, Peegel H, Katta V 1985 Estradiol potentiation of gonadotropin-releasing hormone responsiveness in the anterior pituitary is mediated by an increase in gonadotropin-releasing hormone receptors. Am J Obstet Gynecol 151:534–540
Kaiser UB, Jakubowiak A, Steinberger A, Chin WW 1993 Regulation of rat pituitary gonadotropin-releasing hormone receptor mRNA levels in vivo and in vitro. Endocrinology 133:931–934
Cowley MA, Rao A, Wright PJ, Illing N, Millar RP, Clarke IJ 1998 Evidence for differential regulation of multiple transcripts of the gonadotropin releasing hormone receptor in the ovine pituitary gland: effect of estrogen. Mol Cell Endocrinol 146:141–149
Hamernik DL, Clay CM, Turzillo A, Van Kirk EA, Moss GE 1995 Estradiol increases amounts of messenger ribonucleic acid for gonadotropin-releasing hormone receptors in sheep. Biol Reprod 53:179–185
Turzillo AM, Campion CE, Clay CM, Nett TM 1994 Regulation of gonadotropin-releasing hormone (GnRH) receptor messenger ribonucleic acid and GnRH receptors during the early preovulatory period in the ewe. Endocrinology 135:1353–1358
Turzillo AM, Juengel JL, Nett TM 1995 Pulsatile gonadotropin-releasing hormone (GnRH) increases concentrations of GnRH receptor messenger ribonucleic acid and numbers of GnRH receptors during luteolysis in the ewe. Biol Reprod 53:418–423
Baird DT, Swanston IA, McNeilly AS 1981 Relationship between LH, FSH, and prolactin concentration and the secretion of androgens and estrogens by the preovulatory follicle in the ewe. Biol Reprod 24:1013–1025
Drouva SV, Gorenne I, Laplante E, Rerat E, Enjalbert A, Kordon C 1990 Estradiol modulates protein kinase C activity in the rat pituitary in vivo and in vitro. Endocrinology 126:536–544
Johnson MS, Mitchell R, Thomson FJ 1992 The priming effect of luteinizing hormone-releasing hormone (LHRH) but not LHRH-induced gonadotropin release, can be prevented by certain protein kinase C inhibitors. Mol Cell Endocrinol 85:183–193
Naor Z 1990 Signal transduction mechanisms of Ca2+ mobilizing hormones: the case of gonadotropin-releasing hormone. Endocr Rev 11:326–353
Thomson FJ, Johnson MS, Mitchell R, Wolbers WB, Ison AJ, MacEwan DJ 1993 The differential effects of protein kinase C activators and inhibitors on rat anterior pituitary hormone release. Mol Cell Endocrinol 94:223–234
Thomson FJ, Mitchell R 1993 Differential involvement of phospholipase A2 in phorbol ester-induced luteinizing hormone and growth hormone release from rat anterior pituitary tissue. Mol Cell Endocrinol 95:75–83
Keefer DA, Stumpf WE, Petrusz P 1976 Quantitative autoradiographic assessment of 3H-estradiol uptake in immunocytochemically characterized pituitary cells. Cell Tissue Res 166:25–35
Abraham IM, Han SK, Todman MG, Korach KS, Herbison AE 2003 Estrogen receptor ? mediates rapid estrogen actions on gonadotropin-releasing hormone neurons in vivo. J Neurosci 23:5771–5777
Carlstrom L, Ke ZJ, Unnerstall JR, Cohen RS, Pandey SC 2001 Estrogen modulation of the cyclic AMP response element-binding protein pathway: effects of long-term and acute treatments. Neuroendocrinology 74:227–243
Joels M 1997 Steroid hormones and excitability in the mammalian brain. Front Neuroendocrinol 18:2–48
McEwen BS, Alves SE 1999 Estrogen actions in the central nervous system. Endocr Rev 20:279–307
Wong M, Thompson TL, Moss RL 1996 Nongenomic actions of estrogen in the brain: physiological significance and cellular mechanisms. Crit Rev Neurobiol 10:189–203
Abbas MM, Evans JJ 2000 Regulation of C-fos protein in gonadotrope cells by oxytocin and gonadotropin-releasing hormone. Neuroendocrinology 71:292–300
Tobin VA, Pompolo S, Clarke IJ 2001 The percentage of pituitary gonadotropes with immunoreactive oestradiol receptors increases in the follicular phase of the ovine oestrous cycle. J Neuroendocrinol 13:846–854
Singh P, Muldoon TG 1982 A direct effect on LHRH on anterior pituitary estrogen receptors in the female rat. J Steroid Biochem 16:31–37
Clarke IJ, Cummins JT, de Kretser DM 1983 Pituitary gland function after disconnection from direct hypothalamic influences in the sheep. Neuroendocrinology 36:376–384
Clarke IJ, Cummins JT 1984 Direct pituitary effects of estrogen and progesterone on gonadotropin secretion in the ovariectomized ewe. Neuroendocrinology 39:267–274
Clarke IJ, Cummins JT, Crowder ME, Nett TM 1987 Pituitary receptors for gonadotropin-releasing hormone in relation to changes in pituitary and plasma luteinizing hormone in ovariectomized-hypothalamo pituitary disconnected ewes. I. Effect of changing frequency of gonadotropin-releasing hormone pulses. Biol Reprod 37:749–754
Clarke IJ, Thomas GB, Yao B, Cummins JT 1987 GnRH secretion throughout the ovine estrous cycle. Neuroendocrinology 46:82–88
Clarke IJ, Cummins JT, Jenkin M, Phillips DJ 1989 The oestrogen-induced surge of LH requires a ‘signal’ pattern of gonadotrophin-releasing hormone input to the pituitary gland in the ewe. J Endocrinol 122:127–134
Phillips DJ, Cummins JT, Clarke IJ 1990 Effects of modifying gonadotrophin-releasing hormone input before and after the oestrogen-induced LH surge in ovariectomized ewes with hypothalamo-pituitary disconnection. J Endocrinol 127:223–233
Chaillou E, Baumont R, Tramu G, Tillet Y 2000 Effect of feeding on Fos protein expression in sheep hypothalamus with special reference to the supraoptic and paraventricular nuclei: an immunohistochemical study. Eur J Neurosci 12:4515–4524
Ginty DD, Kornhauser JM, Thompson MA, Bading H, Mayo KE, Takahashi JS, Greenberg ME 1993 Regulation of CREB phosphorylation in the suprachiasmatic nucleus by light and a circadian clock. Science 260:238–241
al Saati T, Clamens S, Cohen-Knafo E, Faye JC, Prats H, Coindre JM, Wafflart J, Caveriviere P, Bayard F, Delsol G 1993 Production of monoclonal antibodies to human estrogen-receptor protein (ER) using recombinant ER (RER). Int J Cancer 55:651–654
Sokal RR, Rohlf FJ 1969 Biometry: principles and practice of statistics in biological research. WH Freeman: San Francisco
Cesnjaj M, Catt KJ, Stojilkovic SS 1994 Coordinate actions of calcium and protein kinase-C in the expression of primary response genes in pituitary gonadotrophs. Endocrinology 135:692–701
Demay F, Tiffoche C, Thieulant ML 1996 Effect of gonadotropin-releasing hormone on estrogen receptor messenger ribonucleic acid level in perifused pituitary cells. Cell Mol Neurobiol 16:397–402
Clarke IJ, Cummins JT, Findlay JK, Burman KJ, Doughton BW 1984 Effects on plasma luteinizing hormone and follicle-stimulating hormone of varying the frequency and amplitude of gonadotropin-releasing hormone pulses in ovariectomized ewes with hypothalamo-pituitary disconnection. Neuroendocrinology 39:214–221
Clarke IJ, Cummins JT 1985 GnRH pulse frequency determines LH pulse amplitude by altering the amount of releasable LH in the pituitary glands of ewes. J Reprod Fertil 73:425–431
Clarke IJ 1993 Variable patterns of gonadotropin-releasing hormone secretion during the estrogen-induced luteinizing hormone surge in ovariectomized ewes. Endocrinology 133:1624–1632
Mercer JE, Phillips DJ, Clarke IJ 1993 Short-term regulation of gonadotropin subunit mRNA levels by estrogen: studies in the hypothalamo-pituitary intact and hypothalamo-pituitary disconnected ewe. J Neuroendocrinol 5:591–596
Clarke IJ, Funder JW, Findlay JK 1982 Relationship between pituitary nuclear oestrogen receptors and the release of LH, FSH and prolactin in the ewe. J Reprod Fertil 64:355–362
Harris TG, Dye S, Robinson JE, Skinner DC, Evans NP 1999 Progesterone can block transmission of the estradiol-induced signal for luteinizing hormone surge generation during a specific period of time immediately after activation of the gonadotropin-releasing hormone surge-generating system. Endocrinology 140:827–834
Kamel F, Krey LC 1982 Gonadal steroid modulation of LHRH-stimulated LH secretion by pituitary cell cultures. Mol Cell Endocrinol 26:151–164
Childs GV, Unabia G, Komak S 2001 Differential expression of estradiol receptors and ? by gonadotropes during the estrous cycle. J Histochem Cytochem 49:665–666
Hille B, Tse A, Tse FW, Bosma MM 1995 Signaling mechanisms during the response of pituitary gonadotropes to GnRH. Recent Prog Horm Res 50:75–95
Tasaka K, Masumoto N, Mizuki J, Ikebuchi Y, Ohmichi M, Kurachi H, Miyake A, Murata Y 1998 Rab3B is essential for GnRH-induced gonadotrophin release from anterior pituitary cells. J Endocrinol 157:267–274
White R, Parker MG 1998 Molecular mechanisms of steroid hormone action. Endocr Relat Cancer 5:1–14
Zhou Y, Watters JJ, Dorsa DM 1996 Estrogen rapidly induces the phosphorylation of the cAMP response element binding protein in rat brain. Endocrinology 137:2163–2166
Kelly MJ, Levin ER 2001 Rapid actions of plasma membrane estrogen receptors. Trends Endocrinol Metab 12:152–156
Clarke IJ 2002 Multifarious effects of estrogen on the pituitary gonadotrope with special emphasis on studies in the ovine species. Arch Physiol Biochem 110:62–73
Webster JC, Pedersen NR, Edwards DP, Beck CA, Miller WL 1995 The 5'-flanking region of the ovine follicle-stimulating hormone-? gene contains six progesterone response elements: three proximal elements are sufficient to increase transcription in the presence of progesterone. Endocrinology 136:1049–1058
Shupnik MA, Weinmann CM, Notides AC, Chin WW 1989 An upstream region of the rat luteinizing hormone ? gene binds estrogen receptor and confers estrogen responsiveness. J Biol Chem 264:80–86
Fan NC, Peng C, Krisinger J, Leung PC 1995 The human gonadotropin-releasing hormone receptor gene: complete structure including multiple promoters, transcription initiation sites, and polyadenylation signals. Mol Cell Endocrinol 107:R1–R8
Razandi M, Alton G, Pedram A, Ghonshani S, Webb P, Levin ER 2003 Identification of a structural determinant necessary for the localization and function of estrogen receptor at the plasma membrane. Mol Cell Biol 23:1633–1646
Razandi M, Pedram A, Park ST, Levin ER 2003 Proximal events in signaling by plasma membrane estrogen receptors. J Biol Chem 278:2701–2712
Hoffman GE, Lyo D 2002 Anatomical markers of activity in neuroendocrine systems: are we all ‘fos-ed out’? J Neuroendocrinol 14:259–268
Szijan I, Parma DL, Engel NI 1992 Expression of c-myc and c-fos protooncogenes in the anterior pituitary gland of the rat: effect of estrogen. Horm Metab Res 24:154–157
Lehman MN, Durham DM, Jansen HT, Adrian B, Goodman RL 1996 Dopaminergic A14/A15 neurons are activated during estradiol negative feedback in anestrous, but not breeding season, ewes. Endocrinology 137:4443–4450
Allen DL, Mitchner NA, Uveges TE, Nephew KP, Khan S, Ben Jonathan N 1997 Cell-specific induction of c-fos expression in the pituitary gland by estrogen. Endocrinology 138:2128–2135
Clarke IJ, Pompolo S, Scott CJ, Rawson JA, Caddy D, Jakubowska AE, Pereira AM 2001 Cells of the arcuate nucleus and ventromedial nucleus of the ovariectomized ewe that respond to oestrogen: a study using Fos immunohistochemistry. J Neuroendocrinol 13:934–941
Shaywitz AJ, Greenberg ME 1999 CREB: a stimulus-induced transcription factor activated by a diverse array of extracellular signals. Annu Rev Biochem 68:821–861
Adams TE, Wagner TO, Sawyer HR, Nett TM 1979 GnRH interaction with anterior pituitary. II. Cyclic AMP as an intracellular mediator in the GnRH activated gonadotroph. Biol Reprod 21:735–747
Kile JP, Nett TM 1994 Differential secretion of follicle-stimulating hormone and luteinizing hormone from ovine pituitary cells following activation of protein kinase A, protein kinase C, or increased intracellular calcium. Biol Reprod 50:49–54
Lonze BE, Ginty DD 2002 Function and regulation of CREB family transcription factors in the nervous system. Neuron 35:605–623(Iain J. Clarke, Vicky A. )
Address all correspondence and requests for reprints to: Professor Iain J. Clarke, Prince Henry’s Institute of Medical Research, P.O. Box 5152, Clayton, Victoria 3168, Australia. E-mail: iain.clarke@phimr.monash.edu.au.
Abstract
Estrogen receptor- (ER) levels in gonadotropes are increased during the follicular phase of the ovine estrous cycle, a time of increased frequency of pulsatile secretion of GnRH and elevated plasma estrogen levels. In the present study, our first aim was to determine which of these factors causes the rise in the number of gonadotropes with ER. Ovariectomized hypothalamo-pituitary disconnected ewes (n = 4–6) received the following treatments: 1) no treatment, 2) injection (im) of 50 μg estradiol benzoate (EB), 3) pulses (300 ng iv) of GnRH every 3 h, 4) GnRH treatment as in group 3 and EB treatment as in group 2, 5) increased frequency of GnRH pulses commencing 20 h before termination, and 6) GnRH treatment as in group 5 with EB treatment. These treatments had predictable effects on plasma LH levels. The number of gonadotropes in which ER was present (by immunohistochemistry) was increased by either GnRH treatment or EB injection, but combined treatment had the greatest effect. Immunohistochemistry was also performed to detect phosphorylated cAMP response element binding protein (pCREB) and Fos protein in gonadotropes. The number of gonadotropes with Fos and with pCREB was increased only in group 6. We conclude that either estrogen or GnRH can up-regulate ER in pituitary gonadotropes. On the other hand, during the period of positive feedback action of estrogen, the appearance of pCREB and Fos in gonadotropes requires the combined action of estrogen and increased frequency of GnRH input. This suggests convergence of signaling for GnRH and estrogen.
Introduction
THE DECLINE IN plasma progesterone levels at the end of the luteal phase of the estrous cycle removes negative feedback on GnRH, allowing for increased frequency of pulsatile discharges of this neurohormone from the hypothalamus (1, 2). This, in turn, increases the frequency of pulsatile discharges of LH from the pituitary gonadotropes leading to growth of ovarian follicles and increased estrogen secretion (3, 4). The increase in circulating levels of estrogen induces a positive feedback effect on the secretion of GnRH and LH, which is a time-delayed mechanism (4). This involves actions of estrogen on cells in the brain as well as action on the pituitary gonadotropes, involving electrophysiological properties and mobilization of secretory granules (3, 5, 6, 7, 8). The specific molecular actions of estrogen that produce the GnRH/LH surge are not well understood, although this is the primary endocrine signal for ovulation. This estrogenic signal that initiates positive feedback occurs in the follicular phase of the normal estrous cycle (4).
Although there is a wealth of literature regarding the subcellular mechanisms for GnRH action in gonadotropes, there is relatively little information on the means by which estrogen acts on these cells. Estrogen regulates the expression of GnRH receptors in vivo (9, 10, 11) and in vitro (12), with an increase occurring during the follicular phase of the estrous cycle in the ewe (13, 14, 15, 16). Nevertheless, responsiveness of the gonadotropes to GnRH does not increase until close to the time of onset of the LH surge (1), suggesting that this increased GnRH receptor expression may be permissive of enhanced secretion at the time of the LH surge, but it is not the cause of a profound increase in responsiveness at the onset of the surge. LH-secreting granules migrate to the periphery of gonadotropes in preparation for the cyclic LH surge (17), which presumably reflects subcellular mechanisms activated by estrogen. This polarization of the granules also occurs in ovariectomized (OVX) ewes given a single im injection of estradiol benzoate (EB) (7).
A series of studies on rat pituitaries show that there is up-regulation of second messenger systems in the proestrous stage of the estrous cycle, suggesting that these factors may be related to the mechanism of positive feedback (18, 19, 20, 21, 22). Because these studies were done with hemipituitaries, however, and various cell types in the rat pituitary express ER (23), it is difficult to ascribe the activities directly to the gonadotrope. One study in particular provides convincing evidence for the action of estrogen to up-regulate protein kinase C in rat pituitaries (18). These data suggest actions of estrogen that may either be genomic or nongenomic, and recent data (24) further indicate rapid effects via phosphorylation of MAPK and phosphorylation of cAMP response element binding protein (CREB). Short- and long-term effects of estrogen on CREB and related proteins, as well as CRE-DNA binding, are seen in various brain regions (25); the former could be caused by rapid nongenomic effects of estrogen (26, 27, 28). Measurement of phosphorylation of CREB is regarded as an index of cellular activation (24, 25). Fos is another marker of cellular response, and GnRH has been shown to increase the percentage of immuno-identified gonadotropes that produce Fos protein (29).
Whereas estrogen up-regulates GnRH receptor levels (see above), we showed that the number of gonadotropes that express estrogen receptor- (ER) is also increased in the follicular phase of the ovine estrous cycle (30). This could be a means of enhancing sensitivity to estrogen at this time. Early work in rat cells (31) suggested that GnRH could increase nuclear ER in pituitary cells. We hypothesized that the increased expression of ER in gonadotropes in the follicular phase of the estrous cycle might be caused by increased frequency of GnRH input, increasing plasma levels of estrogen, or both. To investigate these possibilities, we employed our well characterized model of the OVX hypothalamo-pituitary disconnected (HPD) ewe, which allows the study of the pituitary gland without brain inputs (32, 33). Programmed replacement with pulsatile GnRH and estrogen allows the study of effects on the gonadotrope in vivo that can be correlated with subsequent postmortem analysis (34). We also examined the effects of either increased GnRH pulse frequency and/or estrogen treatment on the appearance of phosphorylated CREB (pCREB) and Fos in the gonadotropes. The treatment paradigm was such that the measurements were taken at the time when positive feedback actions of estrogen are operative (20 h after an im injection of EB) (33).
Materials and Methods
Ethics statement
All experimental procedures were conducted under a protocol approved by the Victorian Institute of Animal Science Animal Ethics Committee.
Sheep
Adult Corriedale female sheep used in this study were kept at the Prince Henry’s Institute facility at Werribee, Melbourne. The animals were bilaterally OVX at least 1 month before the experiment, and the HPD operation was performed as previously described (32). The latter procedure removes all neural inputs to the median eminence, but gonadotropin secretion can be restored by iv administration of GnRH in a pulsatile mode. Immediately after HPD, one external jugular vein was cannulated with a 12-gauge Teflon Dwellcath (Tuta Laboratories Australia Pty. Ltd., Bedford Park, South Australia, Australia) that was connected to a manometer line (Portex Ltd., Hythe, UK) for infusion of GnRH or vehicle (see below). These infusion lines were connected to a Braun Unita 1 infusion pump that could be programmed to deliver 2.25-ml pulses over 6 min as previously described (34). All animals received vaginal progesterone delivery devices (controlled intravaginal drug release) (Lyppard, Cheltenham, Victoria, Australia) for 5 d with GnRH or saline vehicle replacement as detailed below. These were removed at 1600 h on the fifth day and EBL or oil vehicle (1 ml peanut oil) was administered at 1600 h on the sixth day. The animals were killed for recovery of pituitary glands at 1200 h on the seventh day (20 h after EB or oil vehicle).
The animals were assigned to experimental groups as follows (summarized in Fig 1): 1) saline vehicle and oil vehicle (n = 5); 2) saline vehicle and injection (im) of 50 μg EB (Intervet, Baulkham Hills, New South Wales, Australia) (n = 5); 3) pulsatile replacement of GnRH by iv infusion of 300 ng every 3 h with oil vehicle (n = 6); 4) same GnRH regimen as group 3 with EB treatment (n = 4); 5) pulsatile replacement of GnRH by iv infusion of 300 ng every 3 h, switched to hourly pulses of 100 ng between 1600 h on d 6 until 0800 h on d 7, and then switched to 50-ng pulses every 0.5 h between 0800 h on d 7 until 1200 h on d 7 (time of termination), oil vehicle treatment (n = 4); and 6) same GnRH regimen as group 5 with EB treatment (n = 4).
FIG. 1. Experimental design.
The dose of each GnRH pulse was varied in accordance with the interval of delivery, such that the animals received the same amount of GnRH per unit time. In this way, groups 3 and 4 received GnRH replacement at a low frequency, characteristic of that seen during the luteal phase of the estrous cycle (1, 35). Groups 5 and 6 received GnRH replacement where the pulsatile input increased in frequency, as seen in the follicular phase of the estrous cycle. Previous work has shown that the EB treatment of GnRH-replaced OVX-HPD ewes causes a biphasic response in terms of LH secretion, with a time frame similar to that seen in the OVX, hypothalamo-pituitary intact ewe (36, 37). A surge in LH secretion is seen in OVX-HPD ewes between 16 and 20 h after EB treatment, but this reaches only the magnitude of that seen in the hypothalamo-pituitary intact animal if the GnRH pulse frequency is increased to 10-min intervals or if a large bolus of GnRH is delivered at 16 h (36, 37).
Blood samples (5 ml) were taken as follows. On d 6, all animals were sampled over three GnRH pulses delivered at 0900, 1200, and 1500 h; samples were taken 5 min before the delivered pulse and at 7.5, 15, and 30 min after the pulses. For animals in groups 5 and 6, samples were collected over three pulses on d 6 and then over hourly pulses between 0500 and 0700 h and then over half-hourly pulses between 0800 and 1200 h. For sampling around half-hourly pulses on d 7, samples were taken 7.5, 15, and 25 min after the pulses. For groups 1–4, samples were collected at the same times as indicated for groups 5 and 6, even though GnRH or saline pulses were delivered at 3-hourly intervals in these animals; this enabled quantification of changes in total LH released over the last 3 h of the experiment.
The animals were overdosed with sodium pentobarbital (0.5 ml/kg body weight, iv Lethobarb, May and Baker Pty. Ltd., Melbourne, Australia), and the pituitaries were carefully removed from the sella turcica, cut into 3-mm sagittal slices, and placed in a 4% paraformaldehyde in 0.1 M phosphate buffer (PB) (pH 7.4) for 2 d at 4 C. The tissues were then placed in 30% sucrose in 0.1 M PB solution for 5 d at 4 C and then frozen on powdered dry ice and stored at –20 C.
Immunohistochemistry
For all immunohistochemical procedures, antibody incubations were carried out at room temperature unless otherwise specified and were followed by three 10-min washes in 0.05 M PBS. Pituitary sections (7 μm) were cut on a cryostat (Leica CM 1850) at –22 C, thaw-mounted onto Superfrost slides, and stored at –20 C until used for immunohistochemistry. A modified version of a recently published antigen retrieval method was used to detect ER in ovine pituitary tissue (30). This was preceded by incubation with 3% H2O2 in methanol (30 min) to block endogenous peroxidase activity. Slides were then microwaved (four times for 5 min at 700 W) in 0.01 M citrate buffer (pH 6.0) and left to stand in the same buffer for another 20 min. For immunocytochemistry, the sections were incubated in blocking solution (10% normal donkey serum, 0.3% Triton X-100 in 0.1 M PB) for 30 min and then with ER antibody (monoclonal ID5 raised against the N-terminal region of ER, Dako, Carpinteria, CA) at a 1:8 dilution in blocking solution for 72 h at 4 C. The sections were then incubated (1 h) in biotinylated donkey antimouse IgG (Vector Laboratories, Burlingame CA) at a 1:400 dilution and then with Streptavidin horseradish peroxidase (STA-HRP) complex (Amersham, Piscataway, NJ) at a 1:500 dilution for 1 h. After washes in buffer, the sections were preincubated in NiDAB (0.5 mg/ml, 0.02% NH4NiSO4 and 0.026% CoCl2 in 0.1 M PB) for 10 min. This solution was removed and 0.15% H2O2 quickly added for approximately 4 min. The color reaction was monitored and halted by removing the NiDAB solution and washing in 0.1 M PB. Double labeling was then carried out to identify LH cells that also expressed ER. Expression of LH was detected using an anti-LH polyclonal antibody (30) at 1/5000 and visualized with diaminobenzidine (DAB) to give a brown color. Briefly, after immunohistochemistry for ER, sections were washed in 0.05 M PBS and incubated for 30 min in blocking solution containing normal goat serum, then with primary antibody for 48 h at 4 C, followed by incubation (1 h) in secondary antibody (biotinylated goat antirabbit, Vector) at 1/200 dilution. Sections were then incubated with STA-HRP complex for 1 h followed by DAB solution. The color reaction was as described above but without the addition of NH4NiSO4 and CoCl2.
To identify activated LH cells, a separate series of Fos/LH double-labeling and pCREB/LH double-labeling immunohistochemistry was carried out on two pituitary sections per animal per group. Polyclonal rabbit antibodies to Fos protein (AB5, Oncogene Research, Cambridge, MA; 1/1000 dilution) and pCREB (Upstate Biotechnology, Placid, NY; 1/500 dilution) were used as markers for early activation; both were visualized with NiDAB. Sections were washed in 0.05 M PBS followed by pretreatment with 3% H2O2 in methanol, washed in PBS, and then processed for antigen retrieval as described above. After washes in PBS, the sections were incubated with blocking solution containing normal goat serum for 30 min and then in primary antibody for 72 h at 4 C. This was followed by incubation (1 h) in biotinylated goat antirabbit IgG at 1/200 dilution and then in STA-HRP complex at 1/500 dilution for 1 h. After washes in PBS, sections were then incubated with NiDAB solution and color development carried out as above. Double labeling was achieved by incubation with polyclonal rabbit anti-LH antibody visualization with DAB as described above. The antibodies used for labeling Fos (38), pCREB (39), and ER (40) have been characterized previously.
Analysis
Five regions (one in each of the four quadrants and one region in the center) of each coronal anterior pituitary section were examined under x60 magnification. At this magnification, the presence or absence of nuclear and cytoplasmic staining was unambiguous. In each region, 100 cells were counted to give a total of 500 cells per section. Single LH-positive cells and double-labeled LH/ER, LH/Fos, and LH/pCREB cells were scored for each series of sections. Mean percent colocalization was then calculated for each experimental group.
LH RIA
Plasma LH levels were measured using a RIA as previously described (38). In six assays, sensitivity was 0.1 ng/ml, within-assay coefficient of variation was 3.5–9.8% between 3.3 and 6.1 ng/ml, and between-assay coefficient of variation was 12% at 12.7 ng/ml, 10% at 5.0 ng/ml, 7.3% at 6.6 ng/ml, and 6% at 23.3 ng/ml.
Statistical analysis
To determine statistically significant (P < 0.05) differences between means, area under curves (AUCs) of plasma LH concentration vs. time, and for the percentages of gonadotropes that colocalized either ER or Fos or pCREB, two-way ANOVA was performed. Post hoc tests for differences between groups employed the method of least-significant differences. For analysis of percentage data, the arcsin transformation was used as recommended by Sokal and Rohlf (41). In the case where pretreatment values were compared with posttreatment values, paired Student’s t tests were used.
Results
Plasma LH levels were undetectable in the sheep of groups 1 and 2. Analysis of effects of GnRH with either EB or oil treatment was therefore confined to groups 3–6.
Typical examples of the pattern of response to the different GnRH/EB treatments are seen in Fig. 2. LH pulse amplitudes in the animals given GnRH pulses every 3 h plus oil injection (group 3) decreased (not significant) slightly (14%) over the course of the experiment (Table 1), but the LH pulse amplitudes were reduced by 23% (P < 0.01) and broadened (Fig. 2) in animals given the same GnRH treatment with EB injection, (Table 1). When GnRH pulse frequency was increased (with oil treatment; group 5), the LH pulse amplitudes were reduced across the experiment with progressive increase in GnRH/LH pulse frequency (Table 1). In those animals that received EB with increased pulse frequency (group 6), the LH pulse amplitudes were not significantly reduced on hourly pulse frequency, because one animal showed a marked increase in LH pulse amplitude. When group 6 animals were given half-hourly pulses of GnRH, the LH pulse amplitudes were significantly (P < 0.01) reduced (Table 1).
FIG. 2. Examples of plasma LH levels in individual OVX-HPD ewes receiving GnRH on a fixed 3-hourly pulse frequency or increased pulse frequency, with or without EB treatment. Black bars indicate the timing of GnRH pulses, which were either 300 ng (3-hourly) before EB/oil treatment or 50 ng (half-hourly) during the last 3 h of the animal treatments. Examples are given for animals from groups 3 (upper left), 4 (upper right), 5 (lower left), and 6 (lower right).
TABLE 1. Effect of changing GnRH pulse frequency and effect of EB/oil on LH pulse amplitude in OVX-HPD ewes
AUCs of plasma LH concentration vs. time were analyzed for data obtained between 0900 and 1200 h on d 7, because all animals were sampled at the same clock times. As seen in Fig. 2, the LH pulse shapes were different in animals from groups 1–4, and mean AUC data for groups 4–6 are presented in Table 2. Plasma LH was undetectable in groups 1 and 2, so the analysis excluded these. The data show that EB treatment with increased GnRH pulse frequency (group 6) caused greater release of LH than all other treatments. In the partitioning of effects, the increased GnRH pulse frequency produced significantly (P < 0.01) higher values than the fixed-pulse regimen, and EB treatment produced greater values than oil treatment (P < 0.001). The interaction between the two was significant (P < 0.002), indicating that the combination of increased GnRH pulse frequency with EB treatment gave the greatest AUC values.
TABLE 2. Area under plasma LH concentration vs. time curves between 0900 and 1200 h on d 7 (when groups 5 and 6 received half-hourly GnRH pulses)
Figure 3 shows examples of gonadotrope immunostaining with double labeling for ER, Fos, and pCREB. The percentage of gonadotropes that colocalized ER is shown in Table 3. Both GnRH treatment and EB treatment increased the percentage of gonadotropes that colocalized ER (group 1 vs. all other groups). With GnRH treatment alone (groups 2 and 3), the level of colocalization was the same with both pulse paradigms. Three-hourly GnRH pulses plus EB (group 5) had a greater effect than 3-hourly pulses alone (group 2). EB treatment alone (group 4) was more effective in elevating ER in gonadotropes than GnRH given at the 3-hourly pulse frequency but not more effective than increased GnRH pulse frequency alone (group 3). Increased GnRH pulse frequency plus EB (group 6) was more effective than increased GnRH pulse frequency alone (group 3). The statistical analysis showed that there was a statistically significant (P < 0.001) interaction between the GnRH treatment and the EB treatment, which reflects the fact that EB treatment with increased GnRH pulse frequency gave a much bigger response than EB treatment with fixed GnRH pulse frequency.
FIG. 3. Examples of the ER (row 1), Fos (row 2), and pCREB (row 3) immunostaining in gonadotropes (detected with LH immunostaining). Boxed areas in the first and third columns are enlarged in the second and fourth columns. Open arrows show gonadotropes without dual staining, closed arrows show double labeling of gonadotropes with ER (row 1), Fos (row 2), and pCREB. The arrowheads indicate ER and Fos immunoreactive nuclei of pituitary cells that were not immunoreactive for LH.
TABLE 3. Mean (± SEM) percentage of gonadotropes (LH immunopositive) that colocalized ER, Fos, and pCREB as determined by immunostaining
Both the percentage of gonadotropes that colocalized Fos and the percentage that colocalized pCREB were increased only in the group that received increased GnRH pulse frequency and EB (group 6) (Table 3).
Discussion
Using an in vivo model, we have shown that both GnRH and estrogen increase the number of gonadotropes with immunohistochemically identifiable ER. A complex interaction clearly exists between actions of the GnRH receptor and the ER that allows combined regulation of ER, and this may explain the increase in ER in gonadotropes that is seen in the follicular phase of the estrous cycle (30). Furthermore, we have shown that a combination of increased GnRH pulse frequency and estrogen is required to generate Fos and pCREB responses in these cells. Although earlier studies have shown that GnRH stimulates production of immediate early response genes in T3-1 cells and gonadotropes (29, 42), and estrogen induction of Fos has been observed in the pituitary gland (see below), the combined effect on gonadotropes has not been observed previously. An earlier report (43) demonstrated that in vitro treatment of perifused pituitary cell aggregates with GnRH pulses increased the steady-state level of ER mRNA with no increase in LH? mRNA. These data suggested that there may be convergence of signaling of GnRH and nuclear ER in pituitary gonadotropes, and the present data show that this is indeed the case.
The GnRH and estrogen treatments employed in our study caused predictable changes in the plasma levels of LH, highlighting the utility of the model. In particular, an increase in GnRH pulse frequency reduced LH pulse amplitude, as reported previously (44), and EB altered the shape of LH pulses at the time of predicted positive feedback, as shown previously (33). AUCs were the same for 3-hourly and hourly pulses, as predicted from earlier studies, showing the exactitude of frequency and amplitude relationships (45). The present study confirms that, in the OVX ewe without steroid treatment, the inverse relationship between GnRH pulse frequency and pulse amplitude is rigid and is a reflection of the releasable pool of LH within the cell (45). Although estrogen increased AUC units slightly when animals were given EB on a fixed-pulse frequency, most likely because of altered LH pulse shape, this was not statistically significant. The combination of increased pulse frequency and EB treatment elevated total secretion of LH, although surge-like secretion was not seen. This is because the frequency of GnRH input did not reach that seen during the EB-induced surge in the OVX hypothalamo-pituitary intact ewe (46). Higher-frequency GnRH input in combination with EB treatment does cause a full LH surge in this model (36, 37). The combination of increased GnRH pulse frequency and EB did significantly increase AUC, indicating that the animals were sampled at a time when positive feedback mechanisms were operative. This confirms the effect of estrogen to sensitize cells to respond to GnRH (33). These treatments allow dissection of subcellular events within the gonadotropes that relate to the positive feedback phenomenon.
We employed im injection of EB in oil as an effective means of eliciting the positive feedback phenomenon as previously shown in OVX ewes and in OVX-HPD ewes given pulsatile GnRH replacement (47). The positive feedback effect can also be elicited with iv injection of estradiol to OVX ewes (48), which raises plasma estrogen levels for a period of less than 3 h, and nuclear ER binding (measured by nuclear exchange assay) was elevated for up to 6 h. This indicates that the estrogen signal that initiates the positive feedback effect need only be transient and that a time delay is required before the surge in LH secretion occurs. More recently, others (49) have shown that treatment of OVX ewes with estrogen implants for 8 h will cause an LH surge 17–19 h after commencement of treatment. These authors reinforce the notion that the surge-inducing effect of estrogen involves activation (during which time, estrogen activates mechanisms that will cause a surge with a time delay) and transmission (which is not dependent upon the presence of estrogen). It was also shown in cultured rat pituitary cells by Kamel and Krey (50) that enhancement of LH release by estrogen continues beyond the time that estrogen is withdrawn. It is not known whether liganded ERs are required to be present at the time of the surge, but it is clear that the surge-generating action of estrogen occurs at a time before the surge. Presumably, the delay interval allows a sequence of intracellular signaling that is not well understood. In this paper, we show that activation of at least two transcription factors (Fos and pCREB) require increased pulsatile GnRH input as well as estrogen input. Additional use of this model with biochemical analysis of signaling systems will allow greater resolution of this unique phenomenon.
We previously showed that the number of gonadotropes expressing ER is increased in the follicular phase of the estrous cycle (30), but it was not known whether this was because of increased estrogen levels and/or increased GnRH pulse frequency. The number of gonadotropes that also stain for ER/ER? also increases at proestrus in the rat (51). The present work provides novel in vivo data to suggest that both factors are involved. EB treatment alone caused a significant increase in ER as did 3-hourly pulses of GnRH, but the former was more effective. With increased GnRH pulse frequency in the absence of an EB effect, there was no significant increase in the number of gonadotropes in which immunostaining of ER was seen, suggesting that this manipulation does not lead to an increase in ER. Nevertheless, during the period when a positive feedback effect of EB is operative in our model (33), a combination of EB and increased GnRH pulse frequency increased the percentage of ER-positive gonadotropes to the greatest extent. Thus, the combination is the most effective means of up-regulating ER in this cell type. The follicular phase of the estrous cycle is characterized by increased GnRH/LH pulse frequency (1, 2, 35) as well as increasing levels of estrogen (17). Based on the data of the present study, the increase in gonadotropes that are immunopositive for ER at this time (30) is most likely caused by a combination of these two factors.
Statistical analysis of the effect of GnRH pulse frequency and EB treatment on ER in gonadotropes showed a highly significant interaction between these two parameters, which indicates convergence between estrogen action and GnRH action. Earlier work from various labs and in the ovine model (9) has clearly shown that estrogen up-regulates GnRH receptors, and the present data clearly show that GnRH up-regulates ER. The effect of GnRH on ER protein substantiates the results of Demay et al. (43) (see above). The additive effect could involve action of EB via ER present in these cells even after 7 d of pituitary disconnection without replacement (see group 1 data in Table 3) or action by other means. There are few ER? receptors in the ovine pituitary gland (Scott, C., and I. J. Clarke, unpublished data), so the effect is most likely exclusive to mediation via ER.
GnRH action involves a variety of second messenger systems, including the phospholipases C, D, and A2, protein kinase C, and MAPK pathways (20, 21, 22). Phospholipase hydrolysis of the phosphoinositides also plays a pivotal role, and the release of intracellular calcium is important for the secretory mechanism (20, 52). The cAMP/protein kinase A pathway is operative within gonadotropes, although it has been suggested to be more important for transcriptional regulation than the acute secretory response to GnRH (20). Many of these pathways have also been described in immortalized cell lines. Cell membrane ion channel events are a fundamental component of the secretory response to GnRH (52), and Rab3B appears to be involved (53). Ovine gonadotropes produce a wide range of soluble N-ethyl maleimide-sensitive factor attachment proteins/receptors (8), most likely involved in the exocytosis of gonadotropin-containing vesicles as well. Estrogen can act via ER to directly regulate transcription in gonadotropes (54), and rapid estrogen effects are also possible, involving membrane-associated ER and/or cytoplasmic ER interactions with signaling systems (24, 55, 56). Estrogen sensitization of gonadotropes is a well recognized phenomenon, allowing increased responsiveness to GnRH. This could involve an increase in GnRH receptor number, induced by estrogen but not (at least in the OVX-HPD ewe) by increased GnRH pulse frequency (9, 34). The present model will allow further subcellular dissection of the effects of the combination of estrogen and increased GnRH pulse frequency. A range of other intracellular responses are also seen in the gonadotrope at the time of the estrogen-induced surge in LH secretion (57) that could be the consequence of the genomic action of liganded ER complexes.
Estrogen can act directly on gonadotropes to regulate transcription of the gonadotropin subunits and the GnRH receptor. At least in the case of the ovine FSH?, the promoter region contains sites for the binding of progesterone receptor and activator protein 1 (AP-1) and a site for estrogen-dependent repression (58). The rat LH? gene has an imperfect palindromic sequence in the promoter region that confers estrogen responsiveness (59) allowing direct estrogen action. Similar information is lacking for the ovine LH? gene and the common -subunit, but available evidence shows that estrogen negatively regulates the transcription of all of the gonadotropin subunit genes in the pituitary of the EB-treated OVX ewe over the time course relevant to this study (47). This study also showed that LH? gene expression was not reduced over the period of acute negative and positive feedback after EB treatment of OVX-HPD ewes that were receiving hourly GnRH pulses, presumably because GnRH stimulation was maintained, whereas it declined in the negative-feedback phase in the EB-treated OVX ewe. Steady-state mRNA levels for the GnRH receptor, the -subunit, and LH? (but not FSH?) are increased in the follicular phase of the ovine estrous cycle (13). The human GnRH receptor promoter has a variety of cis-acting regulatory sequences including sites for CRE and AP-1 (60). Estrogen significantly up-regulates GnRH receptor transcription in EB-treated, OVX-HPD ewes receiving constant GnRH input, indicating direct pituitary action (34). Thus, genomic signaling of estrogen can regulate the transcription of various genes within the gonadotrope, and up-regulation of ER by both estrogen and GnRH as shown in this study could facilitate the combined effects of the two hormones. As indicated above, effects on the secretion of LH probably involve a range of signaling systems.
It is also possible that convergence of estrogen and GnRH signaling occurs via membrane-associated ER and/or cytoplasmic ER interactions with signaling systems. In other words, estrogen can act to rapidly phosphorylate CREB and/or MAPK (24), and these could converge on GnRH signaling systems. Recent evidence suggests that ER may associate with calveolin to facilitate membrane binding and signaling via G proteins (61, 62). This can facilitate transactivation of the epidermal growth factor receptor by liganded ER (62). A similar scenario may exist for the interaction of ER and the G protein-coupled GnRH receptor. Such convergence of postreceptor signaling might explain why the increase in GnRH pulse frequency and estrogen act together to activate the gonadotrope or whether effects are solely because of genomic action of liganded ER. More work on this cell type is required before such a mechanism is delineated, and our in vivo HPD model offers a means by which this can be examined in nontransformed gonadotropes.
The appearance of Fos protein is generally regarded as a transient response to a specific activator of cells (63). GnRH produces a Fos response in gonadotropes, and although estrogen has been shown to do the same in a number of different types of cells, there are no published reports detailing the response in gonadotropes. A rapid (within 15 min) response of gonadotropes to GnRH has been seen using Fos as an index (29). An interesting feature of this cell type is that a prolonged Fos response may occur. Thus, Szijan et al. (64) found that a large (250 mg) sc injection of estrogen diundecylenate given to male rats produced elevated c-fos and c-myc mRNA levels in the pituitary gland that were sustained for 16–22 h. Because these authors did not define the cell type and various cell types possess ER in the rat (65), this may or may not relate to gonadotropes. In another study, using 1 μg estradiol given sc to OVX rats, Allen et al. (66) found that c-fos expression was elevated within 2 h and remained elevated for 48 h; this response was localized to lactotropes and folliculostellate cells. Sustained Fos/Fos-related antigen responses are also seen in the dopaminergic cells of the hypothalamus of OVX anestrous ewes chronically treated with estrogen (65). On the other hand, Fos responses to a single im injection of EB cause transient responses in brain cells of the OVX ewe (67). In the present study and with a fixed regimen of GnRH stimulation, the number of gonadotropes that were immunopositive for Fos was similar to untreated gonadotropes. There was no increase in the proportion of gonadotropes with Fos after increased frequency of GnRH input. A transient increase may have occurred earlier in response to the increased pulse frequency, but our design could not test for this. On a fixed GnRH pulse regimen, the number of gonadotropes immunostained for Fos was not increased by EB treatment 20 h earlier, but once again, there could have been an earlier transient response. The most significant feature of the present data is that the combination of increased GnRH pulse frequency and EB treatment significantly increased the number of gonadotropes displaying Fos-immunoreactive nuclei 20 h after the commencement of the combined treatment. This is the period when the positive-feedback effect of estrogen is operative (due to the combined effects of increased GnRH pulse frequency and estrogen), and this is clearly associated with activation of the cells, as indicated by the Fos responses. Because AP-1 binding sites are found in the GnRH receptor promoter (see above), this is at least one gene that could be affected by Fos production. As to how estrogen up-regulates GnRH receptor expression in this experimental model (see above), and during the estrous cycle (13), it is possible that this is via up-regulation of second messenger systems, as indicated above, and Fos.
As with Fos, the level of pCREB was increased only in animals that received increased frequency of GnRH pulses and EB. This could be effected by nongenomic action of estrogen, because rapid effects (within 15 min) of estrogen are seen in the brain for phosphorylation of CREB and MAPK (24). If this involves ER-dependent signaling that is rapid (24), it may involve interaction of the ligand with the receptor in the extranuclear compartment of the cell to phosphorylate signaling molecules directly (55). CREB may act as an integrating point for a range of intracellular signaling systems (68). As indicated above for Fos induction, it is possible that either estrogen or increased GnRH pulse frequency could have caused a transient effect to phosphorylate CREB, but at the time point studied, it was only the combination of the two that produced elevated pCREB production at 20 h. Although estrogen might rapidly phosphorylate CREB through nongenomic mechanisms, chronic estrogen treatment has also been shown to increase pCREB and other components of this system (25), so the increase that we have observed with a combination of increased GnRH pulse frequency could be important for both extranuclear activation of second messenger systems and for genomic effects to regulate transcription. As suggested by Zhou et al. (55), phosphorylation of CREB may allow integration of estrogen signaling and membrane receptor signaling. How this relates to the mechanism of positive feedback remains to be elucidated.
A series of studies by Nett and coworkers (69, 70) show that cAMP production is increased by GnRH treatment in ovine gonadotropes. GnRH stimulates parallel release of LH and cAMP (69). Dibutyryl cAMP was also able to stimulate LH release in their culture system (69). Later work confirmed that dibutyryl cAMP as well as phorbol ester (to stimulate protein kinase C) could stimulate LH and FSH release from ovine pituitary cells in primary culture, but a calcium ionophore selectively stimulated LH release (70). Interestingly, it has been reported that a CRE region is not found on the promoter of the ovine FSH? gene, and it is not known whether such a consensus sequence exists in the ovine LH? or common -subunit gene promoters. As mentioned above, the human GnRH receptor promoter contains a CRE site, so pCREB could be relevant to the transcription of this gene. A range of other second messenger systems (which are activated by estrogen and GnRH; see above) (71) can phosphorylate CREB, leading to effects at the level of transcription.
In summary, we have used a unique in vivo model to elucidate combined effects of increased GnRH pulse frequency and estrogen injection on the levels of ER, Fos, and pCREB in the gonadotropes of the ovine pituitary gland. The data show that the proportion of cells displaying immunoreactive ER is increased by both increased GnRH pulse frequency and estrogen treatment, with interactive effects. This combined action explains the increased level of ER seen during the follicular phase of the estrous cycle (30). In addition, the number of cells displaying Fos and pCREB immunoreactivity is increased by the combined treatment (increased GnRH pulse frequency and estrogen treatment) but not by either alone. It is also this combined treatment that brings about the greatest increase in sensitivity of gonadotropes to GnRH with respect to LH secretion. Again, this suggests an interaction between the signaling systems for GnRH receptors and ERs.
Acknowledgments
We thank Alix Rao, Karen Briscoe, and Bruce Doughton for technical assistance and the National Institutes of Health for assay reagents.
References
Clarke IJ 1995 Evidence that the switch from negative to positive feedback at the level of the pituitary gland is an important timing event for the onset of the preovulatory surge in LH in the ewe. J Endocrinol 145:271–282
Moenter SM, Caraty A, Locatelli A, Karsch FJ 1991 Pattern of gonadotropin-releasing hormone (GnRH) secretion leading up to ovulation in the ewe: existence of a preovulatory GnRH surge. Endocrinology 129:1175–1182
Clarke IJ 1996 The Hypothalamo-pituitary axis. In: Hillier SJ, Kitchener HD, Neilsen JP, eds. Scientific essentials of reproductive medicine. London: WB Saunders; 120–132
Clarke IJ 1995 The preovulatory LH surge: a case of a neuroendocrine switch. Trends Endocrinol Metab 6:241–247
Cowley MA, Chen C, Clarke IJ 1999 Estrogen transiently increases delayed rectifier, voltage-dependent potassium currents in ovine gonadotropes. Neuroendocrinology 69:254–260
Heyward PM, Clarke IJ 1995 A transient effect of estrogen on calcium currents and electrophysiological responses to gonadotropin-releasing hormone in ovine gonadotropes. Neuroendocrinology 62:543–552
Thomas SG, Clarke IJ 1997 The positive feedback action of estrogen mobilizes LH-containing, but not FSH-containing secretory granules in ovine gonadotropes. Endocrinology 138:1347–1350
Thomas SG, Takahashi M, Neill JD, Clarke IJ 1998 Components of the neuronal exocytotic machinery in the anterior pituitary of the ovariectomised ewe and the effects of oestrogen in gonadotropes as studied with confocal microscopy. Neuroendocrinology 67:244–259
Clarke IJ, Cummins JT, Crowder ME, Nett TM 1988 Pituitary receptors for gonadotropin-releasing hormone in relation to changes in pituitary and plasma gonadotropins in ovariectomized hypothalamo/pituitary-disconnected ewes. II. A marked rise in receptor number during the acute feedback effects of estradiol. Biol Reprod 39:349–354
Emons G, Hoffmann HG, Brack C, Ortmann O, Sturm R, Ball P, Knuppen R 1988 Modulation of gonadotropin-releasing hormone receptor concentration in cultured female rat pituitary cells by estradiol treatment. J Steroid Biochem 31:751–756
Menon M, Peegel H, Katta V 1985 Estradiol potentiation of gonadotropin-releasing hormone responsiveness in the anterior pituitary is mediated by an increase in gonadotropin-releasing hormone receptors. Am J Obstet Gynecol 151:534–540
Kaiser UB, Jakubowiak A, Steinberger A, Chin WW 1993 Regulation of rat pituitary gonadotropin-releasing hormone receptor mRNA levels in vivo and in vitro. Endocrinology 133:931–934
Cowley MA, Rao A, Wright PJ, Illing N, Millar RP, Clarke IJ 1998 Evidence for differential regulation of multiple transcripts of the gonadotropin releasing hormone receptor in the ovine pituitary gland: effect of estrogen. Mol Cell Endocrinol 146:141–149
Hamernik DL, Clay CM, Turzillo A, Van Kirk EA, Moss GE 1995 Estradiol increases amounts of messenger ribonucleic acid for gonadotropin-releasing hormone receptors in sheep. Biol Reprod 53:179–185
Turzillo AM, Campion CE, Clay CM, Nett TM 1994 Regulation of gonadotropin-releasing hormone (GnRH) receptor messenger ribonucleic acid and GnRH receptors during the early preovulatory period in the ewe. Endocrinology 135:1353–1358
Turzillo AM, Juengel JL, Nett TM 1995 Pulsatile gonadotropin-releasing hormone (GnRH) increases concentrations of GnRH receptor messenger ribonucleic acid and numbers of GnRH receptors during luteolysis in the ewe. Biol Reprod 53:418–423
Baird DT, Swanston IA, McNeilly AS 1981 Relationship between LH, FSH, and prolactin concentration and the secretion of androgens and estrogens by the preovulatory follicle in the ewe. Biol Reprod 24:1013–1025
Drouva SV, Gorenne I, Laplante E, Rerat E, Enjalbert A, Kordon C 1990 Estradiol modulates protein kinase C activity in the rat pituitary in vivo and in vitro. Endocrinology 126:536–544
Johnson MS, Mitchell R, Thomson FJ 1992 The priming effect of luteinizing hormone-releasing hormone (LHRH) but not LHRH-induced gonadotropin release, can be prevented by certain protein kinase C inhibitors. Mol Cell Endocrinol 85:183–193
Naor Z 1990 Signal transduction mechanisms of Ca2+ mobilizing hormones: the case of gonadotropin-releasing hormone. Endocr Rev 11:326–353
Thomson FJ, Johnson MS, Mitchell R, Wolbers WB, Ison AJ, MacEwan DJ 1993 The differential effects of protein kinase C activators and inhibitors on rat anterior pituitary hormone release. Mol Cell Endocrinol 94:223–234
Thomson FJ, Mitchell R 1993 Differential involvement of phospholipase A2 in phorbol ester-induced luteinizing hormone and growth hormone release from rat anterior pituitary tissue. Mol Cell Endocrinol 95:75–83
Keefer DA, Stumpf WE, Petrusz P 1976 Quantitative autoradiographic assessment of 3H-estradiol uptake in immunocytochemically characterized pituitary cells. Cell Tissue Res 166:25–35
Abraham IM, Han SK, Todman MG, Korach KS, Herbison AE 2003 Estrogen receptor ? mediates rapid estrogen actions on gonadotropin-releasing hormone neurons in vivo. J Neurosci 23:5771–5777
Carlstrom L, Ke ZJ, Unnerstall JR, Cohen RS, Pandey SC 2001 Estrogen modulation of the cyclic AMP response element-binding protein pathway: effects of long-term and acute treatments. Neuroendocrinology 74:227–243
Joels M 1997 Steroid hormones and excitability in the mammalian brain. Front Neuroendocrinol 18:2–48
McEwen BS, Alves SE 1999 Estrogen actions in the central nervous system. Endocr Rev 20:279–307
Wong M, Thompson TL, Moss RL 1996 Nongenomic actions of estrogen in the brain: physiological significance and cellular mechanisms. Crit Rev Neurobiol 10:189–203
Abbas MM, Evans JJ 2000 Regulation of C-fos protein in gonadotrope cells by oxytocin and gonadotropin-releasing hormone. Neuroendocrinology 71:292–300
Tobin VA, Pompolo S, Clarke IJ 2001 The percentage of pituitary gonadotropes with immunoreactive oestradiol receptors increases in the follicular phase of the ovine oestrous cycle. J Neuroendocrinol 13:846–854
Singh P, Muldoon TG 1982 A direct effect on LHRH on anterior pituitary estrogen receptors in the female rat. J Steroid Biochem 16:31–37
Clarke IJ, Cummins JT, de Kretser DM 1983 Pituitary gland function after disconnection from direct hypothalamic influences in the sheep. Neuroendocrinology 36:376–384
Clarke IJ, Cummins JT 1984 Direct pituitary effects of estrogen and progesterone on gonadotropin secretion in the ovariectomized ewe. Neuroendocrinology 39:267–274
Clarke IJ, Cummins JT, Crowder ME, Nett TM 1987 Pituitary receptors for gonadotropin-releasing hormone in relation to changes in pituitary and plasma luteinizing hormone in ovariectomized-hypothalamo pituitary disconnected ewes. I. Effect of changing frequency of gonadotropin-releasing hormone pulses. Biol Reprod 37:749–754
Clarke IJ, Thomas GB, Yao B, Cummins JT 1987 GnRH secretion throughout the ovine estrous cycle. Neuroendocrinology 46:82–88
Clarke IJ, Cummins JT, Jenkin M, Phillips DJ 1989 The oestrogen-induced surge of LH requires a ‘signal’ pattern of gonadotrophin-releasing hormone input to the pituitary gland in the ewe. J Endocrinol 122:127–134
Phillips DJ, Cummins JT, Clarke IJ 1990 Effects of modifying gonadotrophin-releasing hormone input before and after the oestrogen-induced LH surge in ovariectomized ewes with hypothalamo-pituitary disconnection. J Endocrinol 127:223–233
Chaillou E, Baumont R, Tramu G, Tillet Y 2000 Effect of feeding on Fos protein expression in sheep hypothalamus with special reference to the supraoptic and paraventricular nuclei: an immunohistochemical study. Eur J Neurosci 12:4515–4524
Ginty DD, Kornhauser JM, Thompson MA, Bading H, Mayo KE, Takahashi JS, Greenberg ME 1993 Regulation of CREB phosphorylation in the suprachiasmatic nucleus by light and a circadian clock. Science 260:238–241
al Saati T, Clamens S, Cohen-Knafo E, Faye JC, Prats H, Coindre JM, Wafflart J, Caveriviere P, Bayard F, Delsol G 1993 Production of monoclonal antibodies to human estrogen-receptor protein (ER) using recombinant ER (RER). Int J Cancer 55:651–654
Sokal RR, Rohlf FJ 1969 Biometry: principles and practice of statistics in biological research. WH Freeman: San Francisco
Cesnjaj M, Catt KJ, Stojilkovic SS 1994 Coordinate actions of calcium and protein kinase-C in the expression of primary response genes in pituitary gonadotrophs. Endocrinology 135:692–701
Demay F, Tiffoche C, Thieulant ML 1996 Effect of gonadotropin-releasing hormone on estrogen receptor messenger ribonucleic acid level in perifused pituitary cells. Cell Mol Neurobiol 16:397–402
Clarke IJ, Cummins JT, Findlay JK, Burman KJ, Doughton BW 1984 Effects on plasma luteinizing hormone and follicle-stimulating hormone of varying the frequency and amplitude of gonadotropin-releasing hormone pulses in ovariectomized ewes with hypothalamo-pituitary disconnection. Neuroendocrinology 39:214–221
Clarke IJ, Cummins JT 1985 GnRH pulse frequency determines LH pulse amplitude by altering the amount of releasable LH in the pituitary glands of ewes. J Reprod Fertil 73:425–431
Clarke IJ 1993 Variable patterns of gonadotropin-releasing hormone secretion during the estrogen-induced luteinizing hormone surge in ovariectomized ewes. Endocrinology 133:1624–1632
Mercer JE, Phillips DJ, Clarke IJ 1993 Short-term regulation of gonadotropin subunit mRNA levels by estrogen: studies in the hypothalamo-pituitary intact and hypothalamo-pituitary disconnected ewe. J Neuroendocrinol 5:591–596
Clarke IJ, Funder JW, Findlay JK 1982 Relationship between pituitary nuclear oestrogen receptors and the release of LH, FSH and prolactin in the ewe. J Reprod Fertil 64:355–362
Harris TG, Dye S, Robinson JE, Skinner DC, Evans NP 1999 Progesterone can block transmission of the estradiol-induced signal for luteinizing hormone surge generation during a specific period of time immediately after activation of the gonadotropin-releasing hormone surge-generating system. Endocrinology 140:827–834
Kamel F, Krey LC 1982 Gonadal steroid modulation of LHRH-stimulated LH secretion by pituitary cell cultures. Mol Cell Endocrinol 26:151–164
Childs GV, Unabia G, Komak S 2001 Differential expression of estradiol receptors and ? by gonadotropes during the estrous cycle. J Histochem Cytochem 49:665–666
Hille B, Tse A, Tse FW, Bosma MM 1995 Signaling mechanisms during the response of pituitary gonadotropes to GnRH. Recent Prog Horm Res 50:75–95
Tasaka K, Masumoto N, Mizuki J, Ikebuchi Y, Ohmichi M, Kurachi H, Miyake A, Murata Y 1998 Rab3B is essential for GnRH-induced gonadotrophin release from anterior pituitary cells. J Endocrinol 157:267–274
White R, Parker MG 1998 Molecular mechanisms of steroid hormone action. Endocr Relat Cancer 5:1–14
Zhou Y, Watters JJ, Dorsa DM 1996 Estrogen rapidly induces the phosphorylation of the cAMP response element binding protein in rat brain. Endocrinology 137:2163–2166
Kelly MJ, Levin ER 2001 Rapid actions of plasma membrane estrogen receptors. Trends Endocrinol Metab 12:152–156
Clarke IJ 2002 Multifarious effects of estrogen on the pituitary gonadotrope with special emphasis on studies in the ovine species. Arch Physiol Biochem 110:62–73
Webster JC, Pedersen NR, Edwards DP, Beck CA, Miller WL 1995 The 5'-flanking region of the ovine follicle-stimulating hormone-? gene contains six progesterone response elements: three proximal elements are sufficient to increase transcription in the presence of progesterone. Endocrinology 136:1049–1058
Shupnik MA, Weinmann CM, Notides AC, Chin WW 1989 An upstream region of the rat luteinizing hormone ? gene binds estrogen receptor and confers estrogen responsiveness. J Biol Chem 264:80–86
Fan NC, Peng C, Krisinger J, Leung PC 1995 The human gonadotropin-releasing hormone receptor gene: complete structure including multiple promoters, transcription initiation sites, and polyadenylation signals. Mol Cell Endocrinol 107:R1–R8
Razandi M, Alton G, Pedram A, Ghonshani S, Webb P, Levin ER 2003 Identification of a structural determinant necessary for the localization and function of estrogen receptor at the plasma membrane. Mol Cell Biol 23:1633–1646
Razandi M, Pedram A, Park ST, Levin ER 2003 Proximal events in signaling by plasma membrane estrogen receptors. J Biol Chem 278:2701–2712
Hoffman GE, Lyo D 2002 Anatomical markers of activity in neuroendocrine systems: are we all ‘fos-ed out’? J Neuroendocrinol 14:259–268
Szijan I, Parma DL, Engel NI 1992 Expression of c-myc and c-fos protooncogenes in the anterior pituitary gland of the rat: effect of estrogen. Horm Metab Res 24:154–157
Lehman MN, Durham DM, Jansen HT, Adrian B, Goodman RL 1996 Dopaminergic A14/A15 neurons are activated during estradiol negative feedback in anestrous, but not breeding season, ewes. Endocrinology 137:4443–4450
Allen DL, Mitchner NA, Uveges TE, Nephew KP, Khan S, Ben Jonathan N 1997 Cell-specific induction of c-fos expression in the pituitary gland by estrogen. Endocrinology 138:2128–2135
Clarke IJ, Pompolo S, Scott CJ, Rawson JA, Caddy D, Jakubowska AE, Pereira AM 2001 Cells of the arcuate nucleus and ventromedial nucleus of the ovariectomized ewe that respond to oestrogen: a study using Fos immunohistochemistry. J Neuroendocrinol 13:934–941
Shaywitz AJ, Greenberg ME 1999 CREB: a stimulus-induced transcription factor activated by a diverse array of extracellular signals. Annu Rev Biochem 68:821–861
Adams TE, Wagner TO, Sawyer HR, Nett TM 1979 GnRH interaction with anterior pituitary. II. Cyclic AMP as an intracellular mediator in the GnRH activated gonadotroph. Biol Reprod 21:735–747
Kile JP, Nett TM 1994 Differential secretion of follicle-stimulating hormone and luteinizing hormone from ovine pituitary cells following activation of protein kinase A, protein kinase C, or increased intracellular calcium. Biol Reprod 50:49–54
Lonze BE, Ginty DD 2002 Function and regulation of CREB family transcription factors in the nervous system. Neuron 35:605–623(Iain J. Clarke, Vicky A. )