Bipotential Effects of Estrogen on Growth Hormone Synthesis and Storage in Vitro
http://www.100md.com
内分泌学杂志 2005年第4期
Department of Neurobiology and Developmental Sciences College of Medicine, University of Arkansas for Medical Science, Little Rock, Arkansas 72205
Address all correspondence and requests for reprints to: Gwen V. Childs, Ph.D., Professor and Chair, Department of Neurobiology and Developmental Sciences, College of Medicine, 4301 West Markham, University of Arkansas for Medical Science, Little Rock, Arkansas 72205. E-mail: childsgwenv@uams.edu.
Abstract
Increased pulses of serum GH coincide with rising estrogens during the reproductive cycle, suggesting estrogen regulation. However, there is lack of agreement about estrogen’s direct effects on the pituitary. Pituitaries from cycling female rats were dispersed and plated for 24 h in defined media containing vehicle or 0.001–250 nM 17?-estradiol. Estrogen (0.01–10 nM) increased the percentages of GH antigen-bearing cells in the anterior pituitary significantly (1.3- to 1.6-fold) and 0.01–1 nM concentrations also stimulated significant increases in GH mRNA-bearing cells and in the integrated OD for GH mRNA. However, 100–250 nM either had no effect or, inhibitory effects on the area of label for GH mRNA. To test estrogen’s effects on expression of GHRH receptors, cultures were stimulated with biotinylated analogs of GHRH and target cells detected by affinity cytochemistry. Estrogen increased GHRH target cells in populations from rats in all stages of the cycle tested. Basal expression of GHRH target cells declined at metestrus. Cultures treated with 0–1 nM estrogen were then dual labeled for bio-GHRH followed by immunolabeling for GH with the antirabbit IgG-ImmPRESS peroxidase polymer. Over 98% of GH cells bound GHRH and 90–96% of GHRH-bound cells contained GH in all treatment groups. Thus, low concentrations of estrogen may stimulate expression of more cells with GH proteins, biotinylated GHRH binding sites, and GH mRNA, whereas high concentrations have no effect, or may reduce GH mRNA. These bipotential effects may help explain the different findings reported in the literature.
Introduction
GENDER DIFFERENCES IN the expression of anterior pituitary (AP) GH have been described for several decades, suggesting gonadal steroid regulation of GH synthesis or secretion (1, 2). In 1965, Frantz and Rabkin (3) reported a 2-fold rise in serum GH across the menstrual cycle. A later study by Faria et al. (4) used a specific GH immunoradiometric assay to assay changes in serum GH in young women every 10 min for 24 h. In comparing women in different phases of the cycle, they discovered significant 2-fold elevations of GH in the late follicular phase of the normal menstrual cycle. Recently, Oveson et al. (5) reported a longitudinal analysis (within the same menstrual cycle of each subject) that demonstrated a rise in GH secretory pulse frequency and amplitude in the preovulatory phase concomitantly with a rise in serum estradiol (5).
Similar cyclic changes in GH expression have been reported for other species. In sheep, there are increases in serum GH and GH mRNA in the late follicular phase (6). A concomitant surge in GH also occurs at the time of the spontaneous or estradiol-induced LH surge in the ewe (7, 8). In ovariectomized or intact primates, estradiol stimulated increases in GH and IGF-1 (9, 10). Our laboratory reported a decline in expression of GH mRNA (detected by percentages of GH cells). The reduction reached a nadir at metestrus and was followed by a rise that peaked on the morning of diestrus and proestrus (11), coinciding with the expected rise in serum estrogen.
Most recently, a study of aromatase knockout female mice by Yan et al. (12) showed low expression of GHRH receptors, GH, and pit-1 mRNA. Estrogen replacement, in vivo, elevated mRNAs for GHRH receptors, pit-1, and GH 1.8-, 1.9-, and 1.6-fold, respectively. In contrast, Lam et al. (13) reported that injections of 25 μg/kg·d estrogen given to ovariectomized rats for 5 d, decreased both GH mRNA and GHRH receptor mRNA (10 d post surgery). Thus, whereas most evidence points to a coincident rise in estrogen and GH, this rise may depend on the physiological state of the animal as well as the concentration of estrogen.
A review of the literature indicates a similar lack of agreement about whether or not estrogen regulates GH cells at the level of the pituitary (14, 15, 16, 17, 18, 19). Some reports show enhancing effects of estradiol on pituitary GH and GH secretion (15, 16) in vitro. However, others have found no enhancing effects on GH (17), or mixed effects, depending on the environment (10, 14, 18, 19). Our previous studies showed bipotential effects of estrogen on GnRH receptors that were dose dependent and also dependent on the stage of the cycle being studied (20). We hypothesized that the differences in the reports may be partly due to the varying concentrations of estrogen used in the studies.
The objective of these studies was to determine whether estrogen had an enhancing effect on the expression of GH mRNA, antigens, or binding for GHRH, in vitro. This report presents the results of these studies showing that low concentrations of estrogen are stimulatory, but higher concentrations are not. Thus, some of the discordant results in the literature may reflect the concentrations used by a given study.
Materials and Methods
Animals and collection of pituitaries
Female Sprague Dawley rats (250 g; 3–5 months old) were used in these studies after 10 d acclimation. They were given water and food ad libitum, and the stage of their cycle was monitored daily by vaginal smears. Rats were used if they had exhibited two consecutive, normal estrous cycles. Diestrous rats were chosen for the estrogen treatment studies, because their pituitary cells exhibited the highest percentages of estrogen-receptive cells (21, 22, 23). However, when we learned that percentages of GHRH-receptive cells were increased by estrogen, estrous, and metestrous rats were also collected for focused studies of biotinylated GHRH binding. Proestrus was not chosen because previous studies had shown that estrogen down-regulated GnRH receptor expression if given to proestrous rats (20).
The rats were anesthetized with sodium pentobarbital within a few minutes of removal from their cage. When they were fully asleep, they were killed with a guillotine, and the pituitary was rapidly removed and placed in DMEM containing 0.1% BSA. The protocol was approved annually by the Animal Care and Use committee because it was conducted in accordance with the Guide for the Care and Use of Laboratory Animals, 1996, National Academy of Science.
The dissociation protocol has been described previously (11, 24, 25, 26, 27). After plating for 1 h on poly-D-lysine-coated coverslips in 24-well trays, the cells were grown for 15–18 h in DMEM supplemented with transferrin, insulin, and sodium selenite (made from a 10x stock solution; Sigma, St. Louis, MO).
Estrogen treatment
A full-dose range was used, with the water soluble, tissue grade 17?-estradiol from Sigma. The compound could be dissolved directly in the DMEM plus supplements (insulin, transferrin, BSA, and sodium selenite), which served as the vehicle control in pilot studies, we tested different times of estrogen treatment of up to 48 h. We found no significant changes with times in culture ranging from 15 h (overnight) to 48 h. Most of the following experiments used 24 h; however, estrogen was also an effective stimulator of GHRH-receptive cells in 15–18 h.
Immunolabeling protocols for GH antigens
Labeling for GH was initially done with 1:110,000–1:126,000 dilution of antirat GH (Hormone Distribution Office, National Institutes of Health) and the Dako (Carpinteria, CA) rapid kit with includes biotinylated antirabbit IgG and streptavidin peroxidase (25). All percentages graphed in the dose-response curves reported in this study are from counts made of fields labeled with this single labeling protocols, with nickel-intensified diaminobenzidine as the peroxidase substrate (11, 24, 25, 26, 27). In absorption controls, anti-GH was absorbed with 10–100 ng/ml GH in either single or dual labeling protocols and the labeling was abolished. Omitting the anti-GH or the antirabbit IgG resulted in no labeling.
After most of the study was completed, a new protocol for immunolabeling was developed to detect cells that coexpress GHRH receptors and GH antigens with the use of the ImmPRESS peroxidase micropolymer linked to antirabbit IgG (Vector Laboratories, Burlingame, CA). This protocol required 1:183,000–1:200,500 dilutions of antirat GH. It will be described in the last section of this presentation, which focuses on biotinylated GHRH.
Protocol for the detection of GH mRNA
For the in situ hybridization, the cells were fixed for 30 min in 2% glutaraldehyde, followed by four 15-min washes in phosphate buffer containing 4.5% sucrose. The protocol has been described previously (11, 24, 28, 29). The antisense and sense oligonucleotide probes for GH mRNA were made by http://www.GeneDetect.com. The antisense probe hybridizes against nucleotides 64–101 of the rat GH gene mRNA (GenBank accession no. U62779). This probe was conjugated to biotin with GreenStar technology that produces a high yield of labeled probes with at least 10 reporter molecules/probe (http://www.GeneDetect.com). This biotin-labeled antisense probe for GH mRNA produced labeling in 24–30% of AP cells in control diestrous or proestrous fields, values that were comparable to those reported previously (11). Optimal dilutions of this probe initially were 8–10 ng/ml.
Controls included fields exposed to biotin-labeled probes for the sense sequence of GH mRNA, or fields exposed to the vehicle instead of the biotinylated antisense sequence. In addition, excess unlabeled probe was added with the labeled probe. Fields showed that the unlabeled probe competed successfully for mRNA sites and abolished labeling if added in an excess of 80 x the labeled probe. These controls are illustrated in Fig. 1.
FIG. 1. In situ hybridization controls in which 10 ng/ml biotinylated antisense oligonucleotide probe for GH mRNA is added with or without 80x excess unlabeled antisense probe. A, Field from estrous rats showing strong labeling for GH mRNA with the probe from http://www.GeneDetect.com. Representative GH cells are noted by a G. B, Parallel culture treated with the same probe competing with excess unlabeled antisense probe. The labeling has been completely abolished. Bar, 25 μm.
Image analysis and densitometry
The full dose-response study was run on three separate groups of animals (two to three animals per group). A post hoc Power Analysis was done with pilot studies, to establish the number of replicates needed. When applied to the changes in percentages of GH cells with estrogen treatment, the analysis indicated that, for a significant 1.6-fold change (SD± 6; P < .05), three replicate experiments would give us a power of 0.90, which is greater than that needed (0.80). A significant 1.5-fold change and three replicates (SD± 6, P < .05) gives a power of 0.82 (http://calculators.stat.ucla.edu/powercalc/normal/n-2-unequal/n-2-uneq-var-power.php; http://www.health.ucalgary.ca/rollin/stats/ssize/n2.html).
Each of the three experiments produced three coverslips bearing cells exposed to a given concentration of estradiol. Before the image analysis program was applied, 18 high-resolution (1600 x 2400) images/experimental group (equally sampling all nine coverslips and the three experiments) were digitized with the use of a SPOT camera (Diagnostic Instruments, Sterling Heights, MI), the x40 objective, and the same lighting and condenser settings. The fields were chosen systematically and included two fields near the center of the culture, and one in each of four quadrants to the upper and lower left and right of center. The camera field display in the microscope oculars was used to prevent overlap of fields. An entire group of experiments was photographed in the same time period, with identical settings.
The BioQuant NovaPrime Image analysis system (Nashville, TN) designed for use with Windows XP was used to detect area and integrated OD of label on these photographs. After a uniform background reading was taken from a blank area in the field, two threshold levels were set. The first threshold detected all cells by their density, including the labeled cells. The second detected only the label. A macro was designed to calculate the percentage of total area assayed that contained label. Limitations in a size threshold automatically excluded readings from debris or red blood cells. The BioQuant NovaPrime also calculated integrated OD measurements for the label that integrated the density of the pixels that detected label x the area covered by the label. The formula used by the system uses the "sum of the negative log 10 of the intensity of the foreground pixels divided by the intensity of the background pixels" (BioQuant user’s manual). It is often used to compare intensity as well as size in micrometers.
To actually do the measurements, each photograph was visualized at high resolution and two to three nonoverlapping regions per photograph were delineated by topography. Thus, from the 18 photographs per dose per experimental group, at least 36 groups of cells were measured. The area measurements allowed us to monitor the total cell areas measured. There was a tendency toward an increase in total cells assayed (based on an increase in plating density in fields treated with 10 or 100 pM estradiol). However, when the total cell areas measured at each dose were averaged (n = 3 experiments), the differences between the concentrations were not significant. Also important is the fact that cultures treated with the highest concentrations of estradiol (100 and 250 nM) remained as well populated as those treated with vehicle alone. Thus, the analysis of cell areas showed that any gains or losses in label area were not due to overall changes in cell area.
Detection of GHRH-receptive cells with a biotinylated analog
The biotinylated ligand detection system has been used since 1983 (26, 27), and the protocol for biotinylated GHRH has been validated previously (25). Biotinylated GHRH is detected only after binding to living, physiologically active cells. Tests of the ligand have shown that optimal labeling of all cells (saturation of binding sites) is seen after 10 min in 1 nM of biotinylated GHRH, which is biologically active (25). Tests of specificity (25) have demonstrated that omission of the biotinylated GHRH prevented labeling for the ligand in either single or dual-labeling protocol. Unlabeled GHRH, GnRH or CRH were used to compete with biotinylated GHRH for tissue receptors. One hundred-fold excess of unlabeled GHRH successfully blocked binding sites and reduced labeling. CRH or GnRH had no blocking effects on the labeling for the biotinylated GHRH (25). Tests also showed that the biotinylated analog was potent in that it stimulated the release of GH (25).
To detect the hormone content of the GHRH target cells, a new dual immunolabeling protocol for GH was run with the use of the ImmPRESS reagent (Vector Laboratories) containing peroxidase-labeled horse antirabbit IgG. This system relies on a novel method of polymerizing the reporter enzyme and then attaching the polymer to a secondary antibody. It thus creates a micropolymer that is more accessible to tissue and results in lower background labeling than with other enzyme complexes. Our tests of labeling for GH showed optimal labeling with ImmPRESS with antibody dilutions in the range of 1:183,000–1:200,500. This protocol also has the advantage that it does not react with any reactive avidin or biotin sites that might remain from the ABC (avidin-biotin complex; Vector Laboratories) detection system used for the biotinylated GHRH.
For the dual labeling, detection of the biotinylated GHRH (bio-GHRH) requires fixation, 3- to 5-min treatment in 0.3% hydrogen peroxide and treatment with the same blocking reagent to be used in the ImmPRESS reaction at the end of the protocol (10 mM phosphate buffer containing 2.5% normal horse serum; this comes with the ImmPRESS kit components). This new ABC-ImmPRESS dual labeling protocol was applied to cells from three groups of diestrous rats, treated with 0.01, 0.1, and 1 nM estrogen for 24 h. After treatment with biotinylated GHRH and fixation, the biotinylated GHRH binding sites are then detected with ABC elite peroxidase system (Vector Laboratories) according to kit instructions (25). After detection of the ABC peroxidase is completed with the use of nickel-intensified (blue-black) diaminobenzidine, the cells are washed in phosphate buffer and moved to the blocking solution again (10 mM phosphate and 2.5% normal horse serum) for 20 min. The blocking solution is then replaced with antirat GH serum diluted 1:200, 500 in the same blocking solution. The cells are incubated for 30 min at 37 C. They are then washed three times in phosphate buffer, and the ImmPRESS solution is applied (220–250 μl or five drops/well), at room temperature for 20 min. The cells are then washed in phosphate buffer (three changes), and the peroxidase is detected by amber diaminobenzidine (25, 26, 27). After washing, they are dehydrated and mounted on slides (a coverslip is placed over the round coverslip, cell side up). Controls involved the omission of the primary antibody, omission of the ImmPRESS solution, or absorption of the primary antibody with 10–100 ng/ml rat GH. The absorption controls showed complete neutralization of labeling with 10 or 100 ng GH per milliliter of diluted antibody. Figure 2 illustrates the immunoabsorption controls.
FIG. 2. Immunoabsorption control showing labeling for GH antigens with 1:200,500 dilution of antirat GH (A) on diestrous rat pituitary cells. Representative GH cells are noted by a G. B, Neutralizing effect on labeling if 100 ng/ml GH is added to the antibody overnight (incubation at 4 C). The labeling is abolished by the addition of GH to the anti-GH. Bar, 25 μm.
Statistics
At least three experiments involving two to three diestrous female rats were run for each experimental group. For cell counts and percentages, we sampled at least three coverslips per group and 200 cells per coverslip. The Power Analysis is described in the preceding section. For the cell counts, each experiment produced an average from the counted fields; however, the final data point reported in the results is the average of three to six replicates ± SEM. Either Student’s t test, one-way ANOVA, or two-way ANOVA were run to detect significant differences (P < 0.05), depending on the number of experimental groups. In the case of one-way ANOVA, Fisher’s least significant differences test (LSD) identified the groups that were different.
Results
Changes in percentages of cells with GH antigens or mRNA after estrogen treatment
A 24-h estrogen treatment resulted in a significant increase in the percentages of cells with GH antigens (detected by immunolabeling) or GH mRNA (detected by in situ hybridization) (Fig. 3). However, the increase depended upon a specific range of concentrations. Figure 3 shows the dose-response curve and the changes in percentages of mRNA or antigen-bearing GH cells (±SEM) after 24 h with each concentration. When one-way ANOVA was run followed by the Fisher’s LSD test, significant increases were seen when GH proteins were detected after exposure to 0.01, 0.1, 1, and 10 nM estradiol (0.01 nM, P = 0.028; 0.1 nM, P = 0.001; 1 nM, P < 0.001, 10 nM, P < 0.049). Higher concentrations were not stimulatory. When GH mRNA was detected, 0.01, 0.1, and 1 nM were stimulatory (P = 0.006; P = 0.008, and P = 0.009, respectively).
FIG. 3. Pituitary cell cultures from three separate groups of diestrous female rats (two rats per group) were exposed to a 24-h treatment with 0.001–250 nM estrogen or vehicle. They were then fixed and labeled either for GH antigens or GH mRNA. Counts of three coverslips per group for three groups showed that low concentrations of estrogen increased the percentages of GH antigen or mRNA-bearing cells. Cultures exposed to concentrations higher than 10 nM showed no changes. Stars, Significantly different from vehicle control.
A two-way ANOVA was also conducted on the data in Fig. 3, to learn whether there was an interaction, comparing data from the detection of GH mRNA and antigens. Each set of data included eight groups (vehicle plus seven doses) and there were two data sets (mRNA and antigens). With this 2 x 8 factorial design, a two-way ANOVA showed significant differences within the individual data sets, and it also showed an interaction between the two groups with an F value of 3.739, degrees of freedom (DF) of 7, 32, and P = 0.005.
Further analysis of these two data sets then showed that values for GH antigens or mRNA were similar after exposure to concentrations less than 0.01 nM, or greater than 1 nM. However, exposure to 0.1 and 1 nM estradiol resulted in significantly more GH cells detected by immunolabeling than detected by in situ hybridization (P = 0.017 for 0.1 nM and P = 0.016 for 1 nM).
Image analysis was therefore conducted on the fields labeled for GH mRNA to learn whether there were increases in density of label for mRNA per cell, which would support the enhanced expression of GH proteins. Estrogen (0.1 and 1 nM) stimulation caused significant increases in label area and/or density (P < 0.003). Figure 4 graphs the percentage of cell areas that contained label, and Fig. 5 graphs the changes in average integrated OD. Also shown in Fig. 4 are significant decreases in the percentage of cell areas that contained label in fields exposed to 100 or 250 nM estrogen (P < 0.001).
FIG. 4. Cultures labeled for GH mRNA in the above experiments were analyzed by the BioQuant NovaPrime Image analysis system and the integrated OD of the label for GH mRNA was calculated. This macro integrates density and area of label. Low concentrations of estrogen increased the integrated OD of label for GH mRNA in the range of 0.01–0.1 nM. Higher concentrations had no effect. Star, Significantly different from vehicle control.
FIG. 5. During the analysis described for Fig. 4, the total area of cells measured was calculated along with the total area of the label for GH mRNA. From this, we derived the percent of total cell area that contained label. Figure 5 shows an increase in area of label with concentrations of 0.1 and 1 nM. Comparing Figs. 4 and 5 reveals that the density alone per cell must have contributed to the increase in integrated OD seen with 0.01 nM in Fig. 4. Similarly, an increase in percent of cell areas that contained label is seen after exposure to 1 nM in Fig. 5, without contributing to the integrated OD (Fig. 4). Figure 5 also shows that the percent of cell areas that contained label is significantly reduced in the highest concentrations of estrogen (100 or 250 nM).
Figure 6 illustrates fields labeled for GH antigens in vehicle control and in cultures stimulated overnight with 0.01 and 0.1 nM estrogen. The immunolabel for GH is black (representative cells are labeled G in each field), and unlabeled cells have a clear background density. The figures show the increased percentages of GH protein bearing cells with these low concentrations of estrogen.
FIG. 6. Fields from diestrous rat showing changes in percentages of cells immunolabeled for GH proteins. The label is black, and the unlabeled cells have a density that ranges from clear to light gray. The photographs are darkened to show these cells. In the microscope, the cells are colorless and difficult to see. A, Vehicle control with two labeled GH cells. The increase in percentage of GH cells is seen in panel B, which is from a culture exposed to 0.01 nM estrogen and Fig. 6C, which shows cells exposed to 0.1 nM estrogen. Representative labeled GH cells are shown by a G next to the cell. Bar, 15 μm.
Figure 7 illustrates fields labeled for GH mRNA in vehicle control cultures and cultures exposed to 0.01, 0.1, and 250 nM estrogen. The increased density of label and number of labeled cells are evident with the lower concentrations of estrogen. Representative cells are labeled with a G. Note that the expression of GH mRNA is lower in fields exposed to 250 nM estrogen.
FIG. 7. Fields showing in situ hybridization labeling for GH mRNA. GH mRNA is in black patches filling areas of the cell. Unlabeled cells are seen as colorless or light gray in the microscope, the densities usually reflect packed granules or other organelles. They were photographed so that the unlabeled cells could be seen. Black patches indicating label for GH are seen in several cells in the vehicle control (A). Representative GH cells are indicated by a G. B and C, Cultures exposed to 0.01 nM and 0.1 nM, respectively. The increase in integrated OD and percentages of cells with GH mRNA is evident. D, Reduced labeling after 24 h in 250 nM estrogen.
Estrogen effects on GHRH receptive cells
Biotinylated analogs of GHRH were used to detect GHRH target cells and test estrogen’s effects on somatotrope expression of GHRH receptors. Figure 8 graphs the estrogen-mediated increase in percentages of GHRH target cells from diestrous rats (±SEM). Significant changes were seen with 0.01 nM (P = 0.009); 0.1 nM (P = 0.001) or 1 nM (P < 0.001) estrogen. In addition, one-way ANOVA and Fisher’s LSD test showed that the differences between 0.01 and 0.1 nM (P = 0.027) and 0.1 and 1 nM (P = 0.012) were also significant. The studies were also expanded to include cell populations from estrous and metestrous rats treated with 0.1 nM estrogen; a concentration that had stimulated an increase in GH antigens and mRNA. Figure 9 shows the significant stimulatory actions of estrogen on GHRH target cells from rats in all three stages of the cycle (estrus, P = 0.026; metestrus, P = 0.013; diestrus, P < 0.001). The percentages of GHRH-bound AP cells in metestrous populations were significantly lower than those from diestrous populations (P = 0.017), correlating well with the lower GH mRNA previously reported for this stage of the cycle (11). A two-way ANOVA was also conducted on the data graphed in Fig. 9. This was collected from three different stages of the cycle with or without 0.1 nM estrogen. Although there were significant differences when cells from metestrous rats were compared with those from diestrous or estrous rats, and when each group was stimulated with estrogen, the two-way ANOVA showed no interactions with an F value of 0.608, DF of 2, 12, and P = 0.561.
FIG. 8. Pituitary cultures from diestrous rats were cultured for 24 h in 0.01–1 nM estrogen. The cells were then stimulated with 1 nM biotinylated GHRH for 10 min and fixed in glutaraldehyde, and the biotinylated analog was detected by affinity cytochemistry (ABC) and nickel-intensified diaminobenzidine (black). Figure 8 graphs the significant increase in GHRH target cells with concentrations of estrogen as low as 0.01 nM. Star, Significantly different from vehicle-control animals. V, Vehicle.
FIG. 9. Pituitary cultures from estrous, metestrous, or diestrous rats were cultured for 18 h in vehicle or 0.1 nM estradiol and then exposed to bio-GHRH for 10 min, fixed and labeled for the ligand. There was a significant decrease in the percent of labeled cells exposed to vehicle when metestrous populations were compared with those in estrous or diestrous (open star). Estrogen stimulated more GHRH target cells in all three groups (closed star).
Finally, the cultures were also dual labeled for biotinylated GHRH and GH to learn whether these changes affected coexpression of the receptors by GH cells. The new ImmPRESS peroxidase polymers were used for this protocol. Controls for the new ImmPRESS reagents applied to the detection of GH antigens showed that 10–100 ng/ml GH neutralized labeling for GH in the pituitary population (Fig. 2 illustrates the results of 100 ng/ml absorption).
When we analyzed the dual labeling for the populations described in Fig. 8, 90 ± 4% of GHRH target cells contained GH and more than 98% of GH cells bound the biotinylated analog. These values are similar to those reported previously (25). Although 96 ± 5% of GHRH target cells contained GH after estrogen treatment, the difference between the control and estrogen-treated groups was not significant.
Figure 10 illustrates the dual labeling for biotinylated GHRH and the new ImmPRESS reagents, for GH. Figure 10, A, D, and F, compares low magnification views of fields treated with vehicle (Fig 10A) with fields treated with 0.1 and 1 nM estrogen (Fig 10, B and C, respectively). There are more densely labeled cells in the estrogen-treated cultures. The higher magnifications show the resolution of the labeling for GHRH binding sites as patches at the cell periphery, or sometimes linear arrays in one region. This is typical of labeling for these ligands after 10 min of exposure. The orange labeling for GH fills the cell and is identical with the pattern seen with the Dako streptavidin kit (25).
FIG. 10. Cells from diestrous female rat were labeled for biotinylated GHRH, as described in Fig. 8, followed by immunolabeling for GH with the new ImmPRESS protocol. A–C, From vehicle control fields. D and E, Fields treated with 0.1 nM estrogen. The circled region in panel D is magnified in panel E. F–H, Fields treated with 1 nM estrogen. The circled region in panel F is magnified in panel G. Label for biotinylated GHRH is in black patches, often at one pole of the cell, although a through focus analysis will sometimes show a broader distribution (C). The gray-black labeling may also be in lines or swirls over or in a cell (B) or more diffuse dark patches (C or G). The fields treated with estrogen show a higher proportion of cells densely colabeled for GHRH and GH. Often the labeling has coalesced in a process in one region of the cell showing dense black for GHRH receptors and orange for GH. In other cells, the label for GHRH remains spread at the periphery and the orange label for GH fills the cell (H). A few cells show GHRH binding sites and little if any GH antigens, as if they were differentiating, or as if they had secreted the GH during the stimulatory period. Bar, 30 μm (A, D, and F); bar, 20 μm (B, C, E, G, and H).
Discussion
This study was initiated to learn whether estrogen had direct effects on GH cell functions. It was also prompted by a review of the literature, which suggested lack of agreement about the efficacy of estrogen in the regulation of GH at the level of the pituitary. Whereas it was clear that the rise and fall in GH levels, in vivo, coincided for the most part with changes in serum estrogen (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12), it was still unclear if estrogen regulated somatotropes at the level of the pituitary.
To learn more about potential stimulatory actions of estrogen on pituitary GH, we tested a full-dose range. We wanted to learn whether some of the controversy was due to well-known bipotential effects of estrogen seen in other systems (20, 30, 31, 32, 33, 34, 35). Our results demonstrated that, after in vitro exposure to relatively low concentrations (<0.1 nM) of estrogen, there are increases in the percentages of pituitary cells that express GH antigens, mRNA, or GHRH receptors. Furthermore, the image analyses detected low, concentration-dependent increases in density of label integrated with an expansion in the area of label. Concentrations higher than 10 nM failed to stimulate an increase if applied over a 24-h period. This bipotential effect of estrogen may explain some of the lack of agreement, as will be discussed below.
The previous studies showing an estrogen-mediated effect were focused on secretory responses as end point assays. The present studies are the first to report the fact that estrogens cause a change in numbers of GH cells. Webb et al. (14) reported that 100 pM to 10 nM estradiol given for 4 d increased basal rat pituitary GH, in vitro, but not the somatotrope secretory responses to GHRH. In contrast, Simard et al. (15) reported that 10–250 pM of 17?-estradiol, estrone, and estriol stimulated both basal and GHRH-stimulated GH secretion in female rats, effects that were blocked by the antiestrogen L156758. A 72-h pretreatment with estradiol also stimulated synthesis of GH to 50% above control levels. Tulipano et al. (16) reported that an estrogen receptor modulator LY117018 stimulated GH secretion in rat pituitary cultures. Whereas the incubation period was shorter, the correlative findings in the present study show that 0.01–1 nM estradiol increased the number of somatotropes that express GH proteins and mRNA and GHRH receptors, thus correlating well with the changes in responsiveness observed by others (14, 15, 16).
Previous in vivo studies have not agreed about estrogen’s enhancement of GHRH receptor mRNA (12, 13). Yan et al. (12) reported that estrogen treatment of aromatase knockout mice increased GHRH receptor gene expression in anterior pituitaries. In contrast, Lam et al. (13) showed that 25 μg/kg estradiol inhibited expression of GHRH receptor mRNA in ovariectomized mice. The difference may be due to the animal model or the concentration of estrogen. The present studies support the work of Yan et al. (12). In addition, new findings in this study show that the changes in cells bearing GHRH binding sites mirror those of GH mRNA or proteins in two ways. First, differential counts show a decline in biotinylated GHRH target cells during metestrus that correlates well with the decline in GH mRNA-bearing cells (11). Second, estrogen exposure to cells from estrous, metestrous, and diestrous animals increases the percentage of GHRH target cells. The dual labeling shows that over 90% of these cells store GH. Thus, the cytochemical evidence points to the fact that estrogen stimulation brings about an increase in GH cells that appear fully competent. We base this argument on previous studies that demonstrated the biological potency of biotinylated GHRH as well as the fact that the ligand can only be detected with living, healthy cells (25).
Those workers that used higher concentrations of estrogens reported no stimulatory effects on GH secretory activity. Fukuta and Martin (17) reported that, in rat AP cultures, 24 h in 100 nM 17?-estradiol had no effect on GHRH-induced GH release. The present studies report new information showing that 100 nM estrogen does not stimulate more cells with GH mRNA or antigens. In fact, the densitometric analysis shows a significant reduction in area of label for GH mRNA after exposure to 100 nM.
In addition, our present studies show that 10 nM estradiol did not stimulate more cells with GH mRNA. This agrees with studies of bovine pituitary cultures by Silverman et al. (18), who reported that a 72-h pretreatment with 10 nM estradiol increased the GH secretory response to 0.1–10 nM GHRH, but not the levels of GH mRNA.
Mixed or negative findings were reported by a third group of studies that used longer term treatment periods. Bethea reported no changes in media GH in pituitary cultures obtained from spayed, intact adult, or juvenile female rhesus monkeys after 18 d treatment, in vivo, with 1 pM to 10 nM 17?-estradiol (10). Our studies were limited to 24 h, and it is possible that the longer term effects of even low concentrations of estrogen treatment, in vivo, would have a different outcome on the overall population, especially because estrogen is important in the conversion of somatomammotropes to mammotropes (36, 37, 38, 39).
Another example of a study of long-term treatment, in vitro, was published by Hauspie et al. (19). This group reported a negative effect of 10 nM estradiol on percentages of monohormonal GH cells in long-term cultures (4 wk). However, this same concentration of estradiol prevented the decline in the subset of GH cells that were multihormonal. Whereas we agree that 10 nM is not a strong stimulatory concentration, its capacity to maintain a subset of multihormonal GH cells may provide important clues about estrogen effects on differentiation.
Thus, our studies expand and clarify the reports about direct estrogen actions on pituitary GH secretion. In some cases, the differences can be explained by the concentration of estradiol used. Other differences can be explained by the species or their physiological state.
In addition to reporting changes in GH cell percentages, new findings in the present studies include the fact that the magnitude of the changes in percentages of cells with GH mRNA was not as great as that with GH antigens. This result is comparable with our previous findings (11) in which we showed that the percentages of cells with GH mRNA and GH antigens are similar for all experimental groups tested except the metestrous rat AP population. During metestrus, there is a significant decrease in GH mRNA [detected by both in situ hybridization and ribonuclease protection assays (11)] accompanied by only a slight decline in cells with GH antigens. This agrees with previous studies from Frawley’s laboratory (36, 37) in which they show no differences in GH secretory capacity in a reverse hemolytic plaque assay with the stage of the estrous cycle. Thus, the lower expression of transcripts at metestrus indicates differential regulation of transcripts vs. proteins in somatotropes.
In the present studies, after exposure to 1–10 nM estradiol, percentages of cells with GH antigens continue to rise to between 40 and 45% (a 1.6-fold increase over basal), whereas there are no further increases in cells with GH mRNA. It is possible that higher concentrations of estrogen may down-regulate synthesis of GH mRNA either directly, or indirectly by increasing GH antigens available for negative short-loop feedback (40, 41). Other causes may relate to differential effects of estrogen on transcription vs. translation in subsets of GH cells.
To summarize, the results of these studies have supported the hypothesis that the rise in serum GH (1, 2, 3, 4, 5, 6, 7, 8, 9, 10) and GH mRNA (11) during the cycle could be regulated by estradiol at the level of the pituitary. The controversy about estrogen effects reported in the literature can be explained, in part, on the basis of differences in concentration used. Estrogen-enhancing effects are seen in a relatively narrow range of concentrations. These bipotential effects add information to a body of literature describing similar actions of estrogen in other organs and organ systems (30, 31, 32, 33, 34, 35).
Acknowledgments
The authors would like to acknowledge the Hormone Distribution Program, National Institutes of Health, and A. Parlow for the antiserum to rat GH. We also thank Dr. Brian T. Miller, University of Texas Medical Branch (Galveston, TX), for the biotinylated analogs of GHRH.
References
Giustina A, Veldhuis JD 2000 Pathophysiology of the neuroregulation of growth hormone secretion in experimental animals and the human. Endocr Rev 19:717–797
Hull KL, Harvey S 2002 GH as a co-gonadotropin: the relevance of correlative changes in GH secretion and reproductive state. J Endocrinol 172:1–19
Frantz AG, Rabkin MT 1965 Effects of estrogen and sex difference on secretion of human growth hormone. J Clin Endocrinol Metab 25:1470–1480
Faria ACS, Bekenstein LW, Booth Jr RA, Vaccaro VA, Asplin CM, Veldhuis JD, Thorner MO, Evans WS 1992 Pulsatile growth hormone release in normal women during the menstrual cycle. Clin Endocrinol (Oxf) 36:591–596
Ovesen P, Vahl N, Fisker S, Veldhuis JD, Christiansen JS, Jorgensen JOL 1998 Increased pulsatile, but not basal, growth hormone secretion rates and plasma insulin-like growth factor I levels during the preovulatory interval in normal women. J Clin Endocrinol Metab 83:1662–1667
Landefeld TD, Suttie JM 1989 Changes in messenger ribonucleic acid concentrations and plasma levels of growth hormone during the ovine estrous cycle and in response to exogenous estradiol. Endocrinology 125:1474–1478
Malven PV, Haglof SA, Jiang H 1995 Serum concentrations of luteinizing hormone, growth hormone, and prolactin in untreated and estradiol-treated ovariectomized ewes after immunoneutralization of hypothalamic neuropeptide Y. J Anim Sci 73:2105–2112
Scanlan N, Skinner DC 2002 Estradiol modulation of growth hormone secretion in the ewe: no growth hormone-releasing hormone neurons and few somatotropes express estradiol receptor. Biol Reprod 66:1267–1273
Copeland KC, Johnson, DM, Kuehl TJ, Castracane VD 1984 Estrogen stimulates growth hormone and somatomedin-C in castrate and intact female baboons. J Clin Endocrinol Metab 58:698–703
Bethea CL 1991 Estrogen action on growth hormone in pituitary cell cultures from adult and juvenile macaques. Endocrinology 129:2110–2118
Childs GV, Unabia G, Wu P 2000 Differential expression of GH mRNA by growth hormone cells and gonadotropes in male and cycling female rats. Endocrinology 141:1560–1570
Yan M, Jones MEE, Hernandez M, Liu D, Simpson ER, Chen C 2004 Functional modification of pituitary somatotropes in the aromatase knockout mouse and the effect of estrogen replacement. Endocrinology 145:604–612
Lam KS, Lee MF, Tam SP, Srivastava G 1996 Gene expression of the receptor for growth hormone releasing hormone is physiologically regulated by glucocorticoids and estrogen. Neuroendocrinology 63:475–480
Webb, CB, Szabo M, Frohman LA 1983 Ectopic growth hormone releasing factor and dibutryl cyclic adenosine monophosphate-stimulated growth hormone release in vitro: effects of corticosterone and estradiol. Endocrinology 113:1191–1196
Simard J, Hubert JF, Hosseinzadeh T, Labrie F 1986 Stimulation of growth hormone release and synthesis by estrogens in rat anterior pituitary cells in culture. Endocrinology 119:2004–2011
Tulipano G, Bonfanti C, Poiesi C, Burattin A, Turazzi S, Barone G, Cozzi R, Bollati A, Valle D, Giustina A 2004 Effects of the selective estrogen receptor modulator LY117018 on growth hormone secretion: in vitro studies. Metabolism 53:563–570
Fukuta J, Martin JB 1986 Influence of sex steroid hormones on rat growth hormone-releasing factor and somatostatin in dispersed pituitary cells. Endocrinology 119:2256–2261
Silverman BL, Kaplan SL, Grumbach MM, Miller WL 1988 Hormonal regulation of growth hormone secretion and messenger ribonucleic acid accumulation in cultured bovine pituitary cells. Endocrinology 123:1236–1241
Hauspie, A, Seuntjens E, Vankelecom, H, Denef C 2003 Stimulation of combinatorial expression of prolactin and glycoprotein hormone -Subunit genes by gonadotropin-releasing hormone and estradiol-17? in single rat pituitary cells during aggregate cell culture. Endocrinology 144:388–399
Lloyd JM, Childs GV 1988 Changes in the number of GnRH-receptive cells during the rat estrous cycle: biphasic effects of estradiol. Neuroendocrinology 48:138–146
Kikuta T, Yamamoto K, Namiki H, Hayashi S 1993 Immunocytochemical localization of estrogen receptor in various anterior pituitary hormone cells of adult male and female rats. Acta Histochem Cytochem 26:609–614
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
Childs, GV 2002 Development of gonadotropes may involve cyclic transdifferentiation of growth hormone cells. Arch Physiol Biochem 110:42–49
Childs GV, Unabia G, Rougeau D 1994 Cells that express luteinizing hormone (LH) and follicle stimulating hormone (FSH) ? subunit mRNAs during the estrous cycle: the major contributors contain LH?, FSH? and/or growth hormone. Endocrinology 134:990–998
Childs GV, Unabia G, Miller BT, Collins TJ 1999 Differential expression of prolactin and gonadotropin antigens by growth hormone releasing hormone (GHRH) target cells from male and female rats. J Endocrinol 162:177–197
Childs GV, Naor Z, Hazum E, Tibolt R, Westlund KM, Hancock MB 1983 Localization of biotinylated gonadotropin releasing hormone on pituitary monolayer cells with avidin-biotin peroxidase complexes. J Histochem Cytochem 31:1422–1425
Childs GV, Naor Z, Hazum E, Tibolt R, Westlund KN, Hancock MB 1983 Cytochemical characterization of pituitary target cells for biotinylated gonadotropin releasing hormone. Peptides 4:549–555
Childs GV 1999 In situ hybridization with non-radioactive probes. In: Methods in molecular biology. Vol 123. In situ hybridization protocols. In: Darby IA, ed. Totowa, NJ: Humana Press; 131–141
Childs GV 1996 Simultaneous identification of a specific gene protein product and transcript using combined immunocytochemistry and in situ hybridization with non-radioactive probes. Scanning Microsc Suppl 10:17–26
vom Saal FS, Timms BG, Montano MM, Palanza P, Thayer KA, Nagel SC, Dhar MD, Ganjam VK, Parmigiani S, Welshons WV 1997 Prostate enlargement in mice due to fetal exposure to low doses of estradiol or diethylstilbestrol and opposite effects at high doses. Proc Natl Acad Sci USA 94:2056–2061
Veldhuis JD 1986 Mechanisms subserving the bipotential actions of estrogen on ovarian cells: studies with a selective anti-estrogen, LY156758, and the sparingly metabolizable estrogen agonist, moxestrol. J Steroid Biochem 24:977–982
Veldhuis JD 1985 Bipotential actions of estrogen on progesterone biosynthesis by ovarian cells. II. Relation of estradiol’s stimulatory actions to cholesterol and progestin metabolism in cultured swine granulosa cells. Endocrinology 117:1076–1083
Veldhuis JD 1985 Bipotential actions of estrogen on progesterone biosynthesis by ovarian cells. I. Relation of estradiol’s inhibitory actions to cholesterol and progestin metabolism in cultured swine granulosa cells. Endocrinology 116:1818–1825
Wagner EJ, Ronnekleiv OK, Bosch MA, Kelly MJ 2001 Estrogen biphasically modifies hypothalamic GABAergic function concomitantly with negative and positive control of luteinizing hormone release. J Neurosci 21:2085–2093
Veldhuis JD, Evans WS, Rogol AD, Thorner MO, Stumpf P 1987 Actions of estradiol on discrete attributes of the luteinizing hormone pulse signal in man. Studies in postmenopausal women treated with pure estradiol. J Clin Invest 79:769–776
Kineman RD, Faught WJ, Frawley LS 1992 Steroids can modulate transdifferentiation of prolactin and growth hormone cells in bovine pituitary cultures. Endocrinology 130:3289–3294
Kineman RD, Henricks DM, Faught WJ, Frawley LS 1991 Fluctuations in the proportions of growth hormone- and prolactin-secreting cells during the bovine estrous cycle. Endocrinology 129:1221–1225
Kineman RD, Faught WJ, Frawley LS 1990 Bovine pituitary cells exhibit a unique form of somatotrope secretory heterogeneity. Endocrinology 127: 2229–2235
Frawley LS, Boockfor FR 1991 Mammosomatotropes: presence and functions in normal and neoplastic pituitary tissue. Endocr Rev 12:337–355
Hurley DL, Smith, EP, Reynolds, GA, Veldhuis, JD, Bowers CY, GH releasing peptide-1 treatment for 7 days causes a dose-dependent decrease in GH mRNA but increases GH intron-containing transcripts in rat pituitary. Program of the 86th Annual Meeting of The Endocrine Society, New Orleans, LA, 2004, p 503 (Abstract P3-157)
Agustsson T, Bjomsson BT 2000 Growth hormone inhibits growth hormone secretion from the rainbow trout pituitary in vitro. Comp Biochem Physiol C Toxicol Pharmacol 126:299–303(Gwen V. Childs, Mary Irut)
Address all correspondence and requests for reprints to: Gwen V. Childs, Ph.D., Professor and Chair, Department of Neurobiology and Developmental Sciences, College of Medicine, 4301 West Markham, University of Arkansas for Medical Science, Little Rock, Arkansas 72205. E-mail: childsgwenv@uams.edu.
Abstract
Increased pulses of serum GH coincide with rising estrogens during the reproductive cycle, suggesting estrogen regulation. However, there is lack of agreement about estrogen’s direct effects on the pituitary. Pituitaries from cycling female rats were dispersed and plated for 24 h in defined media containing vehicle or 0.001–250 nM 17?-estradiol. Estrogen (0.01–10 nM) increased the percentages of GH antigen-bearing cells in the anterior pituitary significantly (1.3- to 1.6-fold) and 0.01–1 nM concentrations also stimulated significant increases in GH mRNA-bearing cells and in the integrated OD for GH mRNA. However, 100–250 nM either had no effect or, inhibitory effects on the area of label for GH mRNA. To test estrogen’s effects on expression of GHRH receptors, cultures were stimulated with biotinylated analogs of GHRH and target cells detected by affinity cytochemistry. Estrogen increased GHRH target cells in populations from rats in all stages of the cycle tested. Basal expression of GHRH target cells declined at metestrus. Cultures treated with 0–1 nM estrogen were then dual labeled for bio-GHRH followed by immunolabeling for GH with the antirabbit IgG-ImmPRESS peroxidase polymer. Over 98% of GH cells bound GHRH and 90–96% of GHRH-bound cells contained GH in all treatment groups. Thus, low concentrations of estrogen may stimulate expression of more cells with GH proteins, biotinylated GHRH binding sites, and GH mRNA, whereas high concentrations have no effect, or may reduce GH mRNA. These bipotential effects may help explain the different findings reported in the literature.
Introduction
GENDER DIFFERENCES IN the expression of anterior pituitary (AP) GH have been described for several decades, suggesting gonadal steroid regulation of GH synthesis or secretion (1, 2). In 1965, Frantz and Rabkin (3) reported a 2-fold rise in serum GH across the menstrual cycle. A later study by Faria et al. (4) used a specific GH immunoradiometric assay to assay changes in serum GH in young women every 10 min for 24 h. In comparing women in different phases of the cycle, they discovered significant 2-fold elevations of GH in the late follicular phase of the normal menstrual cycle. Recently, Oveson et al. (5) reported a longitudinal analysis (within the same menstrual cycle of each subject) that demonstrated a rise in GH secretory pulse frequency and amplitude in the preovulatory phase concomitantly with a rise in serum estradiol (5).
Similar cyclic changes in GH expression have been reported for other species. In sheep, there are increases in serum GH and GH mRNA in the late follicular phase (6). A concomitant surge in GH also occurs at the time of the spontaneous or estradiol-induced LH surge in the ewe (7, 8). In ovariectomized or intact primates, estradiol stimulated increases in GH and IGF-1 (9, 10). Our laboratory reported a decline in expression of GH mRNA (detected by percentages of GH cells). The reduction reached a nadir at metestrus and was followed by a rise that peaked on the morning of diestrus and proestrus (11), coinciding with the expected rise in serum estrogen.
Most recently, a study of aromatase knockout female mice by Yan et al. (12) showed low expression of GHRH receptors, GH, and pit-1 mRNA. Estrogen replacement, in vivo, elevated mRNAs for GHRH receptors, pit-1, and GH 1.8-, 1.9-, and 1.6-fold, respectively. In contrast, Lam et al. (13) reported that injections of 25 μg/kg·d estrogen given to ovariectomized rats for 5 d, decreased both GH mRNA and GHRH receptor mRNA (10 d post surgery). Thus, whereas most evidence points to a coincident rise in estrogen and GH, this rise may depend on the physiological state of the animal as well as the concentration of estrogen.
A review of the literature indicates a similar lack of agreement about whether or not estrogen regulates GH cells at the level of the pituitary (14, 15, 16, 17, 18, 19). Some reports show enhancing effects of estradiol on pituitary GH and GH secretion (15, 16) in vitro. However, others have found no enhancing effects on GH (17), or mixed effects, depending on the environment (10, 14, 18, 19). Our previous studies showed bipotential effects of estrogen on GnRH receptors that were dose dependent and also dependent on the stage of the cycle being studied (20). We hypothesized that the differences in the reports may be partly due to the varying concentrations of estrogen used in the studies.
The objective of these studies was to determine whether estrogen had an enhancing effect on the expression of GH mRNA, antigens, or binding for GHRH, in vitro. This report presents the results of these studies showing that low concentrations of estrogen are stimulatory, but higher concentrations are not. Thus, some of the discordant results in the literature may reflect the concentrations used by a given study.
Materials and Methods
Animals and collection of pituitaries
Female Sprague Dawley rats (250 g; 3–5 months old) were used in these studies after 10 d acclimation. They were given water and food ad libitum, and the stage of their cycle was monitored daily by vaginal smears. Rats were used if they had exhibited two consecutive, normal estrous cycles. Diestrous rats were chosen for the estrogen treatment studies, because their pituitary cells exhibited the highest percentages of estrogen-receptive cells (21, 22, 23). However, when we learned that percentages of GHRH-receptive cells were increased by estrogen, estrous, and metestrous rats were also collected for focused studies of biotinylated GHRH binding. Proestrus was not chosen because previous studies had shown that estrogen down-regulated GnRH receptor expression if given to proestrous rats (20).
The rats were anesthetized with sodium pentobarbital within a few minutes of removal from their cage. When they were fully asleep, they were killed with a guillotine, and the pituitary was rapidly removed and placed in DMEM containing 0.1% BSA. The protocol was approved annually by the Animal Care and Use committee because it was conducted in accordance with the Guide for the Care and Use of Laboratory Animals, 1996, National Academy of Science.
The dissociation protocol has been described previously (11, 24, 25, 26, 27). After plating for 1 h on poly-D-lysine-coated coverslips in 24-well trays, the cells were grown for 15–18 h in DMEM supplemented with transferrin, insulin, and sodium selenite (made from a 10x stock solution; Sigma, St. Louis, MO).
Estrogen treatment
A full-dose range was used, with the water soluble, tissue grade 17?-estradiol from Sigma. The compound could be dissolved directly in the DMEM plus supplements (insulin, transferrin, BSA, and sodium selenite), which served as the vehicle control in pilot studies, we tested different times of estrogen treatment of up to 48 h. We found no significant changes with times in culture ranging from 15 h (overnight) to 48 h. Most of the following experiments used 24 h; however, estrogen was also an effective stimulator of GHRH-receptive cells in 15–18 h.
Immunolabeling protocols for GH antigens
Labeling for GH was initially done with 1:110,000–1:126,000 dilution of antirat GH (Hormone Distribution Office, National Institutes of Health) and the Dako (Carpinteria, CA) rapid kit with includes biotinylated antirabbit IgG and streptavidin peroxidase (25). All percentages graphed in the dose-response curves reported in this study are from counts made of fields labeled with this single labeling protocols, with nickel-intensified diaminobenzidine as the peroxidase substrate (11, 24, 25, 26, 27). In absorption controls, anti-GH was absorbed with 10–100 ng/ml GH in either single or dual labeling protocols and the labeling was abolished. Omitting the anti-GH or the antirabbit IgG resulted in no labeling.
After most of the study was completed, a new protocol for immunolabeling was developed to detect cells that coexpress GHRH receptors and GH antigens with the use of the ImmPRESS peroxidase micropolymer linked to antirabbit IgG (Vector Laboratories, Burlingame, CA). This protocol required 1:183,000–1:200,500 dilutions of antirat GH. It will be described in the last section of this presentation, which focuses on biotinylated GHRH.
Protocol for the detection of GH mRNA
For the in situ hybridization, the cells were fixed for 30 min in 2% glutaraldehyde, followed by four 15-min washes in phosphate buffer containing 4.5% sucrose. The protocol has been described previously (11, 24, 28, 29). The antisense and sense oligonucleotide probes for GH mRNA were made by http://www.GeneDetect.com. The antisense probe hybridizes against nucleotides 64–101 of the rat GH gene mRNA (GenBank accession no. U62779). This probe was conjugated to biotin with GreenStar technology that produces a high yield of labeled probes with at least 10 reporter molecules/probe (http://www.GeneDetect.com). This biotin-labeled antisense probe for GH mRNA produced labeling in 24–30% of AP cells in control diestrous or proestrous fields, values that were comparable to those reported previously (11). Optimal dilutions of this probe initially were 8–10 ng/ml.
Controls included fields exposed to biotin-labeled probes for the sense sequence of GH mRNA, or fields exposed to the vehicle instead of the biotinylated antisense sequence. In addition, excess unlabeled probe was added with the labeled probe. Fields showed that the unlabeled probe competed successfully for mRNA sites and abolished labeling if added in an excess of 80 x the labeled probe. These controls are illustrated in Fig. 1.
FIG. 1. In situ hybridization controls in which 10 ng/ml biotinylated antisense oligonucleotide probe for GH mRNA is added with or without 80x excess unlabeled antisense probe. A, Field from estrous rats showing strong labeling for GH mRNA with the probe from http://www.GeneDetect.com. Representative GH cells are noted by a G. B, Parallel culture treated with the same probe competing with excess unlabeled antisense probe. The labeling has been completely abolished. Bar, 25 μm.
Image analysis and densitometry
The full dose-response study was run on three separate groups of animals (two to three animals per group). A post hoc Power Analysis was done with pilot studies, to establish the number of replicates needed. When applied to the changes in percentages of GH cells with estrogen treatment, the analysis indicated that, for a significant 1.6-fold change (SD± 6; P < .05), three replicate experiments would give us a power of 0.90, which is greater than that needed (0.80). A significant 1.5-fold change and three replicates (SD± 6, P < .05) gives a power of 0.82 (http://calculators.stat.ucla.edu/powercalc/normal/n-2-unequal/n-2-uneq-var-power.php; http://www.health.ucalgary.ca/rollin/stats/ssize/n2.html).
Each of the three experiments produced three coverslips bearing cells exposed to a given concentration of estradiol. Before the image analysis program was applied, 18 high-resolution (1600 x 2400) images/experimental group (equally sampling all nine coverslips and the three experiments) were digitized with the use of a SPOT camera (Diagnostic Instruments, Sterling Heights, MI), the x40 objective, and the same lighting and condenser settings. The fields were chosen systematically and included two fields near the center of the culture, and one in each of four quadrants to the upper and lower left and right of center. The camera field display in the microscope oculars was used to prevent overlap of fields. An entire group of experiments was photographed in the same time period, with identical settings.
The BioQuant NovaPrime Image analysis system (Nashville, TN) designed for use with Windows XP was used to detect area and integrated OD of label on these photographs. After a uniform background reading was taken from a blank area in the field, two threshold levels were set. The first threshold detected all cells by their density, including the labeled cells. The second detected only the label. A macro was designed to calculate the percentage of total area assayed that contained label. Limitations in a size threshold automatically excluded readings from debris or red blood cells. The BioQuant NovaPrime also calculated integrated OD measurements for the label that integrated the density of the pixels that detected label x the area covered by the label. The formula used by the system uses the "sum of the negative log 10 of the intensity of the foreground pixels divided by the intensity of the background pixels" (BioQuant user’s manual). It is often used to compare intensity as well as size in micrometers.
To actually do the measurements, each photograph was visualized at high resolution and two to three nonoverlapping regions per photograph were delineated by topography. Thus, from the 18 photographs per dose per experimental group, at least 36 groups of cells were measured. The area measurements allowed us to monitor the total cell areas measured. There was a tendency toward an increase in total cells assayed (based on an increase in plating density in fields treated with 10 or 100 pM estradiol). However, when the total cell areas measured at each dose were averaged (n = 3 experiments), the differences between the concentrations were not significant. Also important is the fact that cultures treated with the highest concentrations of estradiol (100 and 250 nM) remained as well populated as those treated with vehicle alone. Thus, the analysis of cell areas showed that any gains or losses in label area were not due to overall changes in cell area.
Detection of GHRH-receptive cells with a biotinylated analog
The biotinylated ligand detection system has been used since 1983 (26, 27), and the protocol for biotinylated GHRH has been validated previously (25). Biotinylated GHRH is detected only after binding to living, physiologically active cells. Tests of the ligand have shown that optimal labeling of all cells (saturation of binding sites) is seen after 10 min in 1 nM of biotinylated GHRH, which is biologically active (25). Tests of specificity (25) have demonstrated that omission of the biotinylated GHRH prevented labeling for the ligand in either single or dual-labeling protocol. Unlabeled GHRH, GnRH or CRH were used to compete with biotinylated GHRH for tissue receptors. One hundred-fold excess of unlabeled GHRH successfully blocked binding sites and reduced labeling. CRH or GnRH had no blocking effects on the labeling for the biotinylated GHRH (25). Tests also showed that the biotinylated analog was potent in that it stimulated the release of GH (25).
To detect the hormone content of the GHRH target cells, a new dual immunolabeling protocol for GH was run with the use of the ImmPRESS reagent (Vector Laboratories) containing peroxidase-labeled horse antirabbit IgG. This system relies on a novel method of polymerizing the reporter enzyme and then attaching the polymer to a secondary antibody. It thus creates a micropolymer that is more accessible to tissue and results in lower background labeling than with other enzyme complexes. Our tests of labeling for GH showed optimal labeling with ImmPRESS with antibody dilutions in the range of 1:183,000–1:200,500. This protocol also has the advantage that it does not react with any reactive avidin or biotin sites that might remain from the ABC (avidin-biotin complex; Vector Laboratories) detection system used for the biotinylated GHRH.
For the dual labeling, detection of the biotinylated GHRH (bio-GHRH) requires fixation, 3- to 5-min treatment in 0.3% hydrogen peroxide and treatment with the same blocking reagent to be used in the ImmPRESS reaction at the end of the protocol (10 mM phosphate buffer containing 2.5% normal horse serum; this comes with the ImmPRESS kit components). This new ABC-ImmPRESS dual labeling protocol was applied to cells from three groups of diestrous rats, treated with 0.01, 0.1, and 1 nM estrogen for 24 h. After treatment with biotinylated GHRH and fixation, the biotinylated GHRH binding sites are then detected with ABC elite peroxidase system (Vector Laboratories) according to kit instructions (25). After detection of the ABC peroxidase is completed with the use of nickel-intensified (blue-black) diaminobenzidine, the cells are washed in phosphate buffer and moved to the blocking solution again (10 mM phosphate and 2.5% normal horse serum) for 20 min. The blocking solution is then replaced with antirat GH serum diluted 1:200, 500 in the same blocking solution. The cells are incubated for 30 min at 37 C. They are then washed three times in phosphate buffer, and the ImmPRESS solution is applied (220–250 μl or five drops/well), at room temperature for 20 min. The cells are then washed in phosphate buffer (three changes), and the peroxidase is detected by amber diaminobenzidine (25, 26, 27). After washing, they are dehydrated and mounted on slides (a coverslip is placed over the round coverslip, cell side up). Controls involved the omission of the primary antibody, omission of the ImmPRESS solution, or absorption of the primary antibody with 10–100 ng/ml rat GH. The absorption controls showed complete neutralization of labeling with 10 or 100 ng GH per milliliter of diluted antibody. Figure 2 illustrates the immunoabsorption controls.
FIG. 2. Immunoabsorption control showing labeling for GH antigens with 1:200,500 dilution of antirat GH (A) on diestrous rat pituitary cells. Representative GH cells are noted by a G. B, Neutralizing effect on labeling if 100 ng/ml GH is added to the antibody overnight (incubation at 4 C). The labeling is abolished by the addition of GH to the anti-GH. Bar, 25 μm.
Statistics
At least three experiments involving two to three diestrous female rats were run for each experimental group. For cell counts and percentages, we sampled at least three coverslips per group and 200 cells per coverslip. The Power Analysis is described in the preceding section. For the cell counts, each experiment produced an average from the counted fields; however, the final data point reported in the results is the average of three to six replicates ± SEM. Either Student’s t test, one-way ANOVA, or two-way ANOVA were run to detect significant differences (P < 0.05), depending on the number of experimental groups. In the case of one-way ANOVA, Fisher’s least significant differences test (LSD) identified the groups that were different.
Results
Changes in percentages of cells with GH antigens or mRNA after estrogen treatment
A 24-h estrogen treatment resulted in a significant increase in the percentages of cells with GH antigens (detected by immunolabeling) or GH mRNA (detected by in situ hybridization) (Fig. 3). However, the increase depended upon a specific range of concentrations. Figure 3 shows the dose-response curve and the changes in percentages of mRNA or antigen-bearing GH cells (±SEM) after 24 h with each concentration. When one-way ANOVA was run followed by the Fisher’s LSD test, significant increases were seen when GH proteins were detected after exposure to 0.01, 0.1, 1, and 10 nM estradiol (0.01 nM, P = 0.028; 0.1 nM, P = 0.001; 1 nM, P < 0.001, 10 nM, P < 0.049). Higher concentrations were not stimulatory. When GH mRNA was detected, 0.01, 0.1, and 1 nM were stimulatory (P = 0.006; P = 0.008, and P = 0.009, respectively).
FIG. 3. Pituitary cell cultures from three separate groups of diestrous female rats (two rats per group) were exposed to a 24-h treatment with 0.001–250 nM estrogen or vehicle. They were then fixed and labeled either for GH antigens or GH mRNA. Counts of three coverslips per group for three groups showed that low concentrations of estrogen increased the percentages of GH antigen or mRNA-bearing cells. Cultures exposed to concentrations higher than 10 nM showed no changes. Stars, Significantly different from vehicle control.
A two-way ANOVA was also conducted on the data in Fig. 3, to learn whether there was an interaction, comparing data from the detection of GH mRNA and antigens. Each set of data included eight groups (vehicle plus seven doses) and there were two data sets (mRNA and antigens). With this 2 x 8 factorial design, a two-way ANOVA showed significant differences within the individual data sets, and it also showed an interaction between the two groups with an F value of 3.739, degrees of freedom (DF) of 7, 32, and P = 0.005.
Further analysis of these two data sets then showed that values for GH antigens or mRNA were similar after exposure to concentrations less than 0.01 nM, or greater than 1 nM. However, exposure to 0.1 and 1 nM estradiol resulted in significantly more GH cells detected by immunolabeling than detected by in situ hybridization (P = 0.017 for 0.1 nM and P = 0.016 for 1 nM).
Image analysis was therefore conducted on the fields labeled for GH mRNA to learn whether there were increases in density of label for mRNA per cell, which would support the enhanced expression of GH proteins. Estrogen (0.1 and 1 nM) stimulation caused significant increases in label area and/or density (P < 0.003). Figure 4 graphs the percentage of cell areas that contained label, and Fig. 5 graphs the changes in average integrated OD. Also shown in Fig. 4 are significant decreases in the percentage of cell areas that contained label in fields exposed to 100 or 250 nM estrogen (P < 0.001).
FIG. 4. Cultures labeled for GH mRNA in the above experiments were analyzed by the BioQuant NovaPrime Image analysis system and the integrated OD of the label for GH mRNA was calculated. This macro integrates density and area of label. Low concentrations of estrogen increased the integrated OD of label for GH mRNA in the range of 0.01–0.1 nM. Higher concentrations had no effect. Star, Significantly different from vehicle control.
FIG. 5. During the analysis described for Fig. 4, the total area of cells measured was calculated along with the total area of the label for GH mRNA. From this, we derived the percent of total cell area that contained label. Figure 5 shows an increase in area of label with concentrations of 0.1 and 1 nM. Comparing Figs. 4 and 5 reveals that the density alone per cell must have contributed to the increase in integrated OD seen with 0.01 nM in Fig. 4. Similarly, an increase in percent of cell areas that contained label is seen after exposure to 1 nM in Fig. 5, without contributing to the integrated OD (Fig. 4). Figure 5 also shows that the percent of cell areas that contained label is significantly reduced in the highest concentrations of estrogen (100 or 250 nM).
Figure 6 illustrates fields labeled for GH antigens in vehicle control and in cultures stimulated overnight with 0.01 and 0.1 nM estrogen. The immunolabel for GH is black (representative cells are labeled G in each field), and unlabeled cells have a clear background density. The figures show the increased percentages of GH protein bearing cells with these low concentrations of estrogen.
FIG. 6. Fields from diestrous rat showing changes in percentages of cells immunolabeled for GH proteins. The label is black, and the unlabeled cells have a density that ranges from clear to light gray. The photographs are darkened to show these cells. In the microscope, the cells are colorless and difficult to see. A, Vehicle control with two labeled GH cells. The increase in percentage of GH cells is seen in panel B, which is from a culture exposed to 0.01 nM estrogen and Fig. 6C, which shows cells exposed to 0.1 nM estrogen. Representative labeled GH cells are shown by a G next to the cell. Bar, 15 μm.
Figure 7 illustrates fields labeled for GH mRNA in vehicle control cultures and cultures exposed to 0.01, 0.1, and 250 nM estrogen. The increased density of label and number of labeled cells are evident with the lower concentrations of estrogen. Representative cells are labeled with a G. Note that the expression of GH mRNA is lower in fields exposed to 250 nM estrogen.
FIG. 7. Fields showing in situ hybridization labeling for GH mRNA. GH mRNA is in black patches filling areas of the cell. Unlabeled cells are seen as colorless or light gray in the microscope, the densities usually reflect packed granules or other organelles. They were photographed so that the unlabeled cells could be seen. Black patches indicating label for GH are seen in several cells in the vehicle control (A). Representative GH cells are indicated by a G. B and C, Cultures exposed to 0.01 nM and 0.1 nM, respectively. The increase in integrated OD and percentages of cells with GH mRNA is evident. D, Reduced labeling after 24 h in 250 nM estrogen.
Estrogen effects on GHRH receptive cells
Biotinylated analogs of GHRH were used to detect GHRH target cells and test estrogen’s effects on somatotrope expression of GHRH receptors. Figure 8 graphs the estrogen-mediated increase in percentages of GHRH target cells from diestrous rats (±SEM). Significant changes were seen with 0.01 nM (P = 0.009); 0.1 nM (P = 0.001) or 1 nM (P < 0.001) estrogen. In addition, one-way ANOVA and Fisher’s LSD test showed that the differences between 0.01 and 0.1 nM (P = 0.027) and 0.1 and 1 nM (P = 0.012) were also significant. The studies were also expanded to include cell populations from estrous and metestrous rats treated with 0.1 nM estrogen; a concentration that had stimulated an increase in GH antigens and mRNA. Figure 9 shows the significant stimulatory actions of estrogen on GHRH target cells from rats in all three stages of the cycle (estrus, P = 0.026; metestrus, P = 0.013; diestrus, P < 0.001). The percentages of GHRH-bound AP cells in metestrous populations were significantly lower than those from diestrous populations (P = 0.017), correlating well with the lower GH mRNA previously reported for this stage of the cycle (11). A two-way ANOVA was also conducted on the data graphed in Fig. 9. This was collected from three different stages of the cycle with or without 0.1 nM estrogen. Although there were significant differences when cells from metestrous rats were compared with those from diestrous or estrous rats, and when each group was stimulated with estrogen, the two-way ANOVA showed no interactions with an F value of 0.608, DF of 2, 12, and P = 0.561.
FIG. 8. Pituitary cultures from diestrous rats were cultured for 24 h in 0.01–1 nM estrogen. The cells were then stimulated with 1 nM biotinylated GHRH for 10 min and fixed in glutaraldehyde, and the biotinylated analog was detected by affinity cytochemistry (ABC) and nickel-intensified diaminobenzidine (black). Figure 8 graphs the significant increase in GHRH target cells with concentrations of estrogen as low as 0.01 nM. Star, Significantly different from vehicle-control animals. V, Vehicle.
FIG. 9. Pituitary cultures from estrous, metestrous, or diestrous rats were cultured for 18 h in vehicle or 0.1 nM estradiol and then exposed to bio-GHRH for 10 min, fixed and labeled for the ligand. There was a significant decrease in the percent of labeled cells exposed to vehicle when metestrous populations were compared with those in estrous or diestrous (open star). Estrogen stimulated more GHRH target cells in all three groups (closed star).
Finally, the cultures were also dual labeled for biotinylated GHRH and GH to learn whether these changes affected coexpression of the receptors by GH cells. The new ImmPRESS peroxidase polymers were used for this protocol. Controls for the new ImmPRESS reagents applied to the detection of GH antigens showed that 10–100 ng/ml GH neutralized labeling for GH in the pituitary population (Fig. 2 illustrates the results of 100 ng/ml absorption).
When we analyzed the dual labeling for the populations described in Fig. 8, 90 ± 4% of GHRH target cells contained GH and more than 98% of GH cells bound the biotinylated analog. These values are similar to those reported previously (25). Although 96 ± 5% of GHRH target cells contained GH after estrogen treatment, the difference between the control and estrogen-treated groups was not significant.
Figure 10 illustrates the dual labeling for biotinylated GHRH and the new ImmPRESS reagents, for GH. Figure 10, A, D, and F, compares low magnification views of fields treated with vehicle (Fig 10A) with fields treated with 0.1 and 1 nM estrogen (Fig 10, B and C, respectively). There are more densely labeled cells in the estrogen-treated cultures. The higher magnifications show the resolution of the labeling for GHRH binding sites as patches at the cell periphery, or sometimes linear arrays in one region. This is typical of labeling for these ligands after 10 min of exposure. The orange labeling for GH fills the cell and is identical with the pattern seen with the Dako streptavidin kit (25).
FIG. 10. Cells from diestrous female rat were labeled for biotinylated GHRH, as described in Fig. 8, followed by immunolabeling for GH with the new ImmPRESS protocol. A–C, From vehicle control fields. D and E, Fields treated with 0.1 nM estrogen. The circled region in panel D is magnified in panel E. F–H, Fields treated with 1 nM estrogen. The circled region in panel F is magnified in panel G. Label for biotinylated GHRH is in black patches, often at one pole of the cell, although a through focus analysis will sometimes show a broader distribution (C). The gray-black labeling may also be in lines or swirls over or in a cell (B) or more diffuse dark patches (C or G). The fields treated with estrogen show a higher proportion of cells densely colabeled for GHRH and GH. Often the labeling has coalesced in a process in one region of the cell showing dense black for GHRH receptors and orange for GH. In other cells, the label for GHRH remains spread at the periphery and the orange label for GH fills the cell (H). A few cells show GHRH binding sites and little if any GH antigens, as if they were differentiating, or as if they had secreted the GH during the stimulatory period. Bar, 30 μm (A, D, and F); bar, 20 μm (B, C, E, G, and H).
Discussion
This study was initiated to learn whether estrogen had direct effects on GH cell functions. It was also prompted by a review of the literature, which suggested lack of agreement about the efficacy of estrogen in the regulation of GH at the level of the pituitary. Whereas it was clear that the rise and fall in GH levels, in vivo, coincided for the most part with changes in serum estrogen (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12), it was still unclear if estrogen regulated somatotropes at the level of the pituitary.
To learn more about potential stimulatory actions of estrogen on pituitary GH, we tested a full-dose range. We wanted to learn whether some of the controversy was due to well-known bipotential effects of estrogen seen in other systems (20, 30, 31, 32, 33, 34, 35). Our results demonstrated that, after in vitro exposure to relatively low concentrations (<0.1 nM) of estrogen, there are increases in the percentages of pituitary cells that express GH antigens, mRNA, or GHRH receptors. Furthermore, the image analyses detected low, concentration-dependent increases in density of label integrated with an expansion in the area of label. Concentrations higher than 10 nM failed to stimulate an increase if applied over a 24-h period. This bipotential effect of estrogen may explain some of the lack of agreement, as will be discussed below.
The previous studies showing an estrogen-mediated effect were focused on secretory responses as end point assays. The present studies are the first to report the fact that estrogens cause a change in numbers of GH cells. Webb et al. (14) reported that 100 pM to 10 nM estradiol given for 4 d increased basal rat pituitary GH, in vitro, but not the somatotrope secretory responses to GHRH. In contrast, Simard et al. (15) reported that 10–250 pM of 17?-estradiol, estrone, and estriol stimulated both basal and GHRH-stimulated GH secretion in female rats, effects that were blocked by the antiestrogen L156758. A 72-h pretreatment with estradiol also stimulated synthesis of GH to 50% above control levels. Tulipano et al. (16) reported that an estrogen receptor modulator LY117018 stimulated GH secretion in rat pituitary cultures. Whereas the incubation period was shorter, the correlative findings in the present study show that 0.01–1 nM estradiol increased the number of somatotropes that express GH proteins and mRNA and GHRH receptors, thus correlating well with the changes in responsiveness observed by others (14, 15, 16).
Previous in vivo studies have not agreed about estrogen’s enhancement of GHRH receptor mRNA (12, 13). Yan et al. (12) reported that estrogen treatment of aromatase knockout mice increased GHRH receptor gene expression in anterior pituitaries. In contrast, Lam et al. (13) showed that 25 μg/kg estradiol inhibited expression of GHRH receptor mRNA in ovariectomized mice. The difference may be due to the animal model or the concentration of estrogen. The present studies support the work of Yan et al. (12). In addition, new findings in this study show that the changes in cells bearing GHRH binding sites mirror those of GH mRNA or proteins in two ways. First, differential counts show a decline in biotinylated GHRH target cells during metestrus that correlates well with the decline in GH mRNA-bearing cells (11). Second, estrogen exposure to cells from estrous, metestrous, and diestrous animals increases the percentage of GHRH target cells. The dual labeling shows that over 90% of these cells store GH. Thus, the cytochemical evidence points to the fact that estrogen stimulation brings about an increase in GH cells that appear fully competent. We base this argument on previous studies that demonstrated the biological potency of biotinylated GHRH as well as the fact that the ligand can only be detected with living, healthy cells (25).
Those workers that used higher concentrations of estrogens reported no stimulatory effects on GH secretory activity. Fukuta and Martin (17) reported that, in rat AP cultures, 24 h in 100 nM 17?-estradiol had no effect on GHRH-induced GH release. The present studies report new information showing that 100 nM estrogen does not stimulate more cells with GH mRNA or antigens. In fact, the densitometric analysis shows a significant reduction in area of label for GH mRNA after exposure to 100 nM.
In addition, our present studies show that 10 nM estradiol did not stimulate more cells with GH mRNA. This agrees with studies of bovine pituitary cultures by Silverman et al. (18), who reported that a 72-h pretreatment with 10 nM estradiol increased the GH secretory response to 0.1–10 nM GHRH, but not the levels of GH mRNA.
Mixed or negative findings were reported by a third group of studies that used longer term treatment periods. Bethea reported no changes in media GH in pituitary cultures obtained from spayed, intact adult, or juvenile female rhesus monkeys after 18 d treatment, in vivo, with 1 pM to 10 nM 17?-estradiol (10). Our studies were limited to 24 h, and it is possible that the longer term effects of even low concentrations of estrogen treatment, in vivo, would have a different outcome on the overall population, especially because estrogen is important in the conversion of somatomammotropes to mammotropes (36, 37, 38, 39).
Another example of a study of long-term treatment, in vitro, was published by Hauspie et al. (19). This group reported a negative effect of 10 nM estradiol on percentages of monohormonal GH cells in long-term cultures (4 wk). However, this same concentration of estradiol prevented the decline in the subset of GH cells that were multihormonal. Whereas we agree that 10 nM is not a strong stimulatory concentration, its capacity to maintain a subset of multihormonal GH cells may provide important clues about estrogen effects on differentiation.
Thus, our studies expand and clarify the reports about direct estrogen actions on pituitary GH secretion. In some cases, the differences can be explained by the concentration of estradiol used. Other differences can be explained by the species or their physiological state.
In addition to reporting changes in GH cell percentages, new findings in the present studies include the fact that the magnitude of the changes in percentages of cells with GH mRNA was not as great as that with GH antigens. This result is comparable with our previous findings (11) in which we showed that the percentages of cells with GH mRNA and GH antigens are similar for all experimental groups tested except the metestrous rat AP population. During metestrus, there is a significant decrease in GH mRNA [detected by both in situ hybridization and ribonuclease protection assays (11)] accompanied by only a slight decline in cells with GH antigens. This agrees with previous studies from Frawley’s laboratory (36, 37) in which they show no differences in GH secretory capacity in a reverse hemolytic plaque assay with the stage of the estrous cycle. Thus, the lower expression of transcripts at metestrus indicates differential regulation of transcripts vs. proteins in somatotropes.
In the present studies, after exposure to 1–10 nM estradiol, percentages of cells with GH antigens continue to rise to between 40 and 45% (a 1.6-fold increase over basal), whereas there are no further increases in cells with GH mRNA. It is possible that higher concentrations of estrogen may down-regulate synthesis of GH mRNA either directly, or indirectly by increasing GH antigens available for negative short-loop feedback (40, 41). Other causes may relate to differential effects of estrogen on transcription vs. translation in subsets of GH cells.
To summarize, the results of these studies have supported the hypothesis that the rise in serum GH (1, 2, 3, 4, 5, 6, 7, 8, 9, 10) and GH mRNA (11) during the cycle could be regulated by estradiol at the level of the pituitary. The controversy about estrogen effects reported in the literature can be explained, in part, on the basis of differences in concentration used. Estrogen-enhancing effects are seen in a relatively narrow range of concentrations. These bipotential effects add information to a body of literature describing similar actions of estrogen in other organs and organ systems (30, 31, 32, 33, 34, 35).
Acknowledgments
The authors would like to acknowledge the Hormone Distribution Program, National Institutes of Health, and A. Parlow for the antiserum to rat GH. We also thank Dr. Brian T. Miller, University of Texas Medical Branch (Galveston, TX), for the biotinylated analogs of GHRH.
References
Giustina A, Veldhuis JD 2000 Pathophysiology of the neuroregulation of growth hormone secretion in experimental animals and the human. Endocr Rev 19:717–797
Hull KL, Harvey S 2002 GH as a co-gonadotropin: the relevance of correlative changes in GH secretion and reproductive state. J Endocrinol 172:1–19
Frantz AG, Rabkin MT 1965 Effects of estrogen and sex difference on secretion of human growth hormone. J Clin Endocrinol Metab 25:1470–1480
Faria ACS, Bekenstein LW, Booth Jr RA, Vaccaro VA, Asplin CM, Veldhuis JD, Thorner MO, Evans WS 1992 Pulsatile growth hormone release in normal women during the menstrual cycle. Clin Endocrinol (Oxf) 36:591–596
Ovesen P, Vahl N, Fisker S, Veldhuis JD, Christiansen JS, Jorgensen JOL 1998 Increased pulsatile, but not basal, growth hormone secretion rates and plasma insulin-like growth factor I levels during the preovulatory interval in normal women. J Clin Endocrinol Metab 83:1662–1667
Landefeld TD, Suttie JM 1989 Changes in messenger ribonucleic acid concentrations and plasma levels of growth hormone during the ovine estrous cycle and in response to exogenous estradiol. Endocrinology 125:1474–1478
Malven PV, Haglof SA, Jiang H 1995 Serum concentrations of luteinizing hormone, growth hormone, and prolactin in untreated and estradiol-treated ovariectomized ewes after immunoneutralization of hypothalamic neuropeptide Y. J Anim Sci 73:2105–2112
Scanlan N, Skinner DC 2002 Estradiol modulation of growth hormone secretion in the ewe: no growth hormone-releasing hormone neurons and few somatotropes express estradiol receptor. Biol Reprod 66:1267–1273
Copeland KC, Johnson, DM, Kuehl TJ, Castracane VD 1984 Estrogen stimulates growth hormone and somatomedin-C in castrate and intact female baboons. J Clin Endocrinol Metab 58:698–703
Bethea CL 1991 Estrogen action on growth hormone in pituitary cell cultures from adult and juvenile macaques. Endocrinology 129:2110–2118
Childs GV, Unabia G, Wu P 2000 Differential expression of GH mRNA by growth hormone cells and gonadotropes in male and cycling female rats. Endocrinology 141:1560–1570
Yan M, Jones MEE, Hernandez M, Liu D, Simpson ER, Chen C 2004 Functional modification of pituitary somatotropes in the aromatase knockout mouse and the effect of estrogen replacement. Endocrinology 145:604–612
Lam KS, Lee MF, Tam SP, Srivastava G 1996 Gene expression of the receptor for growth hormone releasing hormone is physiologically regulated by glucocorticoids and estrogen. Neuroendocrinology 63:475–480
Webb, CB, Szabo M, Frohman LA 1983 Ectopic growth hormone releasing factor and dibutryl cyclic adenosine monophosphate-stimulated growth hormone release in vitro: effects of corticosterone and estradiol. Endocrinology 113:1191–1196
Simard J, Hubert JF, Hosseinzadeh T, Labrie F 1986 Stimulation of growth hormone release and synthesis by estrogens in rat anterior pituitary cells in culture. Endocrinology 119:2004–2011
Tulipano G, Bonfanti C, Poiesi C, Burattin A, Turazzi S, Barone G, Cozzi R, Bollati A, Valle D, Giustina A 2004 Effects of the selective estrogen receptor modulator LY117018 on growth hormone secretion: in vitro studies. Metabolism 53:563–570
Fukuta J, Martin JB 1986 Influence of sex steroid hormones on rat growth hormone-releasing factor and somatostatin in dispersed pituitary cells. Endocrinology 119:2256–2261
Silverman BL, Kaplan SL, Grumbach MM, Miller WL 1988 Hormonal regulation of growth hormone secretion and messenger ribonucleic acid accumulation in cultured bovine pituitary cells. Endocrinology 123:1236–1241
Hauspie, A, Seuntjens E, Vankelecom, H, Denef C 2003 Stimulation of combinatorial expression of prolactin and glycoprotein hormone -Subunit genes by gonadotropin-releasing hormone and estradiol-17? in single rat pituitary cells during aggregate cell culture. Endocrinology 144:388–399
Lloyd JM, Childs GV 1988 Changes in the number of GnRH-receptive cells during the rat estrous cycle: biphasic effects of estradiol. Neuroendocrinology 48:138–146
Kikuta T, Yamamoto K, Namiki H, Hayashi S 1993 Immunocytochemical localization of estrogen receptor in various anterior pituitary hormone cells of adult male and female rats. Acta Histochem Cytochem 26:609–614
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
Childs, GV 2002 Development of gonadotropes may involve cyclic transdifferentiation of growth hormone cells. Arch Physiol Biochem 110:42–49
Childs GV, Unabia G, Rougeau D 1994 Cells that express luteinizing hormone (LH) and follicle stimulating hormone (FSH) ? subunit mRNAs during the estrous cycle: the major contributors contain LH?, FSH? and/or growth hormone. Endocrinology 134:990–998
Childs GV, Unabia G, Miller BT, Collins TJ 1999 Differential expression of prolactin and gonadotropin antigens by growth hormone releasing hormone (GHRH) target cells from male and female rats. J Endocrinol 162:177–197
Childs GV, Naor Z, Hazum E, Tibolt R, Westlund KM, Hancock MB 1983 Localization of biotinylated gonadotropin releasing hormone on pituitary monolayer cells with avidin-biotin peroxidase complexes. J Histochem Cytochem 31:1422–1425
Childs GV, Naor Z, Hazum E, Tibolt R, Westlund KN, Hancock MB 1983 Cytochemical characterization of pituitary target cells for biotinylated gonadotropin releasing hormone. Peptides 4:549–555
Childs GV 1999 In situ hybridization with non-radioactive probes. In: Methods in molecular biology. Vol 123. In situ hybridization protocols. In: Darby IA, ed. Totowa, NJ: Humana Press; 131–141
Childs GV 1996 Simultaneous identification of a specific gene protein product and transcript using combined immunocytochemistry and in situ hybridization with non-radioactive probes. Scanning Microsc Suppl 10:17–26
vom Saal FS, Timms BG, Montano MM, Palanza P, Thayer KA, Nagel SC, Dhar MD, Ganjam VK, Parmigiani S, Welshons WV 1997 Prostate enlargement in mice due to fetal exposure to low doses of estradiol or diethylstilbestrol and opposite effects at high doses. Proc Natl Acad Sci USA 94:2056–2061
Veldhuis JD 1986 Mechanisms subserving the bipotential actions of estrogen on ovarian cells: studies with a selective anti-estrogen, LY156758, and the sparingly metabolizable estrogen agonist, moxestrol. J Steroid Biochem 24:977–982
Veldhuis JD 1985 Bipotential actions of estrogen on progesterone biosynthesis by ovarian cells. II. Relation of estradiol’s stimulatory actions to cholesterol and progestin metabolism in cultured swine granulosa cells. Endocrinology 117:1076–1083
Veldhuis JD 1985 Bipotential actions of estrogen on progesterone biosynthesis by ovarian cells. I. Relation of estradiol’s inhibitory actions to cholesterol and progestin metabolism in cultured swine granulosa cells. Endocrinology 116:1818–1825
Wagner EJ, Ronnekleiv OK, Bosch MA, Kelly MJ 2001 Estrogen biphasically modifies hypothalamic GABAergic function concomitantly with negative and positive control of luteinizing hormone release. J Neurosci 21:2085–2093
Veldhuis JD, Evans WS, Rogol AD, Thorner MO, Stumpf P 1987 Actions of estradiol on discrete attributes of the luteinizing hormone pulse signal in man. Studies in postmenopausal women treated with pure estradiol. J Clin Invest 79:769–776
Kineman RD, Faught WJ, Frawley LS 1992 Steroids can modulate transdifferentiation of prolactin and growth hormone cells in bovine pituitary cultures. Endocrinology 130:3289–3294
Kineman RD, Henricks DM, Faught WJ, Frawley LS 1991 Fluctuations in the proportions of growth hormone- and prolactin-secreting cells during the bovine estrous cycle. Endocrinology 129:1221–1225
Kineman RD, Faught WJ, Frawley LS 1990 Bovine pituitary cells exhibit a unique form of somatotrope secretory heterogeneity. Endocrinology 127: 2229–2235
Frawley LS, Boockfor FR 1991 Mammosomatotropes: presence and functions in normal and neoplastic pituitary tissue. Endocr Rev 12:337–355
Hurley DL, Smith, EP, Reynolds, GA, Veldhuis, JD, Bowers CY, GH releasing peptide-1 treatment for 7 days causes a dose-dependent decrease in GH mRNA but increases GH intron-containing transcripts in rat pituitary. Program of the 86th Annual Meeting of The Endocrine Society, New Orleans, LA, 2004, p 503 (Abstract P3-157)
Agustsson T, Bjomsson BT 2000 Growth hormone inhibits growth hormone secretion from the rainbow trout pituitary in vitro. Comp Biochem Physiol C Toxicol Pharmacol 126:299–303(Gwen V. Childs, Mary Irut)