Both Estrogen Receptor α and β Stimulate Pituitary GH Gene Expression

Abstract

Although sex steroids have been implicated in the control of mammalian growth, their direct effect on GH synthesis is less clear. The aim of this study was to establish whether estradiol (E₂) directly affects GH synthesis in somatotrophs. Somatotroph GH₃ and MtT/S cells were used as in vitro models. At physiological doses of E₂ stimulation, GH mRNA levels were increased and the ER antagonist ICI 182,780 completely abolished this effect. Estrogen receptor (ER) α– and ERβ-selective agonists, propylpyrazole triol (PPT), and 2,3-bis(4-hydroxyphenyl) propionitrile (DPN), respectively, augmented GH mRNA expression and secretion, whereas E₂ and PPT, but not DPN increased prolactin (PRL) mRNA levels. E₂, PPT, and DPN stimulated expression of the pituitary transcription factor Pou1f1 and increased its binding to the GH promoter. In vivo evidence of E₂ effects on GH synthesis was obtained from the generation of the somatotroph-specific ERα knockout (sERα-KO) mouse model. Basal pituitary GH, PRL, POU1F1, and ERα mRNA expression levels were lower in sERα-KO mice compared with those in controls; whereas ERβ mRNA levels remained unchanged. E₂ and DPN stimulated pituitary GH mRNA expression and serum GH levels in control and sERα-KO ovariectomized mice; however, serum GH levels were unchanged in PPT-treated ovariectomized sERα-KO mice. In these animal models, PRL mRNA levels increased after either E₂ or PPT, but an increase was not seen after DPN treatment. Thus, we propose a mechanism by which estrogen directly regulates somatotroph GH synthesis at a pretranslational level. In contrast to the predominant effect of ERα in the lactotroph, these results support a role for both ERα and ERβ in the transcriptional control of Gh in the somatotroph and illustrate important differences in ER isoform specificity in the anterior pituitary gland.

GH production is mainly regulated by 3 peptide hormones: GHRH, somatostatin, and ghrelin. In addition, sex steroids are important regulators of GH production and action and are thought to act both centrally at the level of the hypothalamus and pituitary gland and peripherally to control human growth. It is generally accepted that of the sex steroids, estrogen plays the major role in regulation of GH production (1). Serum GH levels have been shown to increase during puberty, with estrogen treatment and to correlate with estrogen levels during the menstrual cycle (1, 2). Results from in vitro studies, however, are conflicting with some showing an increase, some showing a decrease, and some showing no effect on GH synthesis after treatment with estrogen (2).

The biological actions of estrogens are mediated by binding to estrogen receptors (ERs), which belong to the nuclear steroid hormone receptor family of transcription factors (3, 4). High levels of ERs are expressed in the pituitary and the hypothalamus (5–9) as 2 distinct isoforms: ERα and ERβ. After ligand binding, ERα and ERβ form homodimers or heterodimers and classically bind to specific estrogen response elements in the promoter region of target genes. Through a variety of mechanisms and interactions with other transcription factors and distinct coactivators and corepressors, these binding events can result in either transactivation or transrepression of target gene expression (10–13). A synergistic effect of ERα and POU1F1 (Pit-1), a pituitary-specific transcription factor, has been shown to increase the Prl gene expression (14). POU1F1 is a member of the POU family of transcriptional activating proteins, which is responsible for pituitary development and hormone production from cells of the somatolactotroph lineage. Containing 2 protein domains (POU-specific and POU-homeo), POU1F1 regulates hormone production by binding to high-affinity sites on the Gh and Prl genes and activates gene transcription.

The aim of this study was to establish whether estrogen affects somatotroph GH production and the mechanism by which this is accomplished. We used 2 different somatotroph cell lines, GH₃ and MtT/S. These cell lines both contain ER isoforms (7, 15, 16) and express GH at high levels (17). Moreover, we extended our in vitro findings to mice by generating a somatotroph-specific ERα knockout (sERα-KO) mouse model. Our findings demonstrate that prolactin (PRL) synthesis in the pituitary lactrotroph is only regulated by ERα, whereas GH from the pituitary somatotroph is regulated by both ERα and ERβ.

Materials and Methods

Cell cultures

GH₃ cells were obtained from American Type Culture Collection, and MtT/S cells were purchased from the Riken Bioresource Center Cell Bank. GH₃ cells were grown in DMEM, supplemented with 10% fetal bovine serum (FBS), 2 mM l-glutamine, and an antibiotic-antimycotic mixture (100 μg/mL penicillin G sodium, 100 μg/mL streptomycin sulfate, and 250 ng/mL amphotericin B) (Invitrogen), and the MtT/S cells were grown in DMEM/F12 culture medium supplemented with 10% FBS, 2.5% horse serum, and an antibiotic-antimycotic mixture, under 5% CO₂ at 37°C. Cells were plated in 6-well plates with 60% to 70% confluence; for MtT/S cells, l-lysine-coated plates were used (Becton Dickinson Labware). The medium was replaced 24 hours later with phenol red–free DMEM (Invitrogen) supplemented with 10% charcoal-stripped FBS and an antibiotic-antimycotic mixture for 24 hours before treatment. Cells were treated with 17β-estradiol (E₂), propylpyrazole triol (PPT), and 2,3-bis(4-hydroxyphenyl) propionitrile (DPN) (Cayman Chemical) and ICI 182,780 (Ascent Scientific).

Animal housing

Mice were maintained with food and water ad libitum in a 14:10-hour light/dark cycle at 24°C in the Broadway Research Building animal facility at the Johns Hopkins University School of Medicine (Baltimore, Maryland). All procedures were approved by the Johns Hopkins University Animal Care and Use Committee.

RNA extraction and quantitative (q) RT-PCR analyses

Cells were washed with ice-cold PBS (pH 7.4) (Invitrogen), and RNA was extracted using the TRIzol/chloroform/2-propanol method (Invitrogen and Fisher Scientific). Pituitary and hypothalamic tissues were homogenized in TRIzol reagent, and total RNA was extracted by the same method. The RNA concentration was measured by an Epoch Multi-Volume Spectrophotometer System (BioTek), and 2 μg of the total RNA was used for the RT reaction with an iScript cDNA Synthesis kit (Bio-Rad Laboratories). Two microliters of the cDNA reaction was used in the presence of 2.5 U of Taq polymerase (Gene Choice), 0.2 μM concentrations of each primer, 0.2 μM concentrations of each nucleotide, and standard reaction buffer (Gene Choice). qPCR analyses were performed using IQ SYBR Green Supermix (Bio-Rad Laboratories) and the IQ5 Multicolor Real-Time PCR Detection System (Bio-Rad Laboratories), according to the manufacturer's protocols. The sequences of the primer sets used for experiments were as follows: ERα, sense 5′-CCAATTCTGACAATCGACGC-3′ and antisense 5′-TCTTATCGATGGTGCATTGGTT-3′ (18); ERβ, sense 5′-GCTGTGATGAACTACAGTGTTCCC-3′ and antisense 5′-TGGACTAGTAACAGGGCTGGCACA-3′) (18); GH, sense 5′-GCTGCAGACTCTCAGACTCCCTGG-3′ and antisense5′-CTGAGAAGCAGAACGCAGCCTG-3′ (19); PRL, sense 5′-ACTAATGACTGCCCCACTTC-3′ and antisense 5′-ATTCCAGGAGTGCACCAAAC-3′ (20); POU1F1, sense 5′-ACTCAGGGTGTGGTCTGGAAACTT-3′ and antisense 5′-ATGTCCACAGCGACAGGACTT CAT-3′; and 18S, sense 5′-TGGTTGATCCTGCCAGTAG-3′ and antisense 5′-CGACCAAAGGAACCATAACT-3′ (21). All PCRs were run for 35 cycles with an annealing temperature of 59°C. A cycle threshold (C_t) value was obtained for each sample. A corrected C_t (ΔC_t) was calculated by subtracting 18S C_t from the corresponding C_t of each unknown sample. Each experiment was repeated with cells from different passages. Relative fold differences of the normalized vehicle controls were calculated using the ΔΔC_t method. In addition, PCR products were also analyzed by agarose gel electrophoresis (data not shown).

DNA extraction and PCR analysis

Three-month-old male Wistar rats were killed by phenobarbital injection and cervical dislocation. Different organ and tissue samples were obtained, and DNA was extracted using the phenol/chloroform/2-propanol method. Samples were normalized to 100 ng of DNA per reaction, and PCR analysis for ERα and ERβ was performed, using Taq DNA polymerase (Choice Taq DNA polymerase; Denville Scientific). The specific primer sets and conditions are listed above.

Western blot analyses

After hormonal treatment, cells were washed with ice-cold PBS, and lysates were prepared using cell lysis buffer (Cell Signaling Technology) supplemented with complete protease inhibitor cocktail (Roche Applied Science). Samples were centrifuged for 5 minutes at 12 000 rpm (clarifying spin), and supernatants were stored at −80°C for further processing. Protein levels for each sample were measured using a BCA protein assay kit (Thermo Fisher Scientific), and the samples were normalized to protein concentration. Proteins were separated by 12% SDS-PAGE and transferred to nitrocellulose membrane for 90 minutes at 500 mA. Membranes were incubated overnight with 5% blotting-grade blocker nonfat dry milk (Bio-Rad Laboratories) in Tris-buffered saline plus 0.1% Tween 20 solution and incubated with specific antibodies. The primary antibodies used were anti-ERα (C1355; rabbit, 1:2000 dilution, 06–935; Millipore Corporation) and anti-ERβ (rabbit, 1:2000 dilution, PA1–311; Affinity Bioreagents). Goat anti-rabbit IgG (H+L)-horseradish peroxidase conjugates (Bio-Rad Laboratories) was used as a secondary antibody for blotting.

Chromatin immunoprecipitation (ChIP) assay followed by qRT-PCR analyses

The ChIP assay was performed using the ChIP-IT Express kit (Active Motif), following the manufacturer's instructions. In brief, MtT/S cells were cultured as described previously and then treated with E₂ for 18 hours. Cells were fixed with 1% formaldehyde in growth medium for 10 minutes at room temperature. After cell and nuclear lysis, the cross-linked DNA was sheared to 200- to 1000-bp lengths by sonication on ice in a Branson Sonifier 250-A (6 pulses of 10-second-long bursts at power output 3; duty cycle, 50%). The sheared cross-linked chromatin was immunoprecipitated (IP) in dilution buffer and incubated overnight at 4°C with 3 μg of rabbit polyclonal Pit-1 antibody (sc-442; Santa Cruz Biotechnology) or 1 μg of normal mouse IgG (sc-2025; Santa Cruz Biotechnology) as a negative control. Protein G magnetic beads (25 μL) were added to the samples to collect the immune complexes. After the beads were washed, protein-DNA complexes were eluted and reverse cross-linked at 95°C for 15 minutes by incubation in ChIP elution buffer and then treated with proteinase K. After the genomic DNA was treated with proteinase K stop solution, it was used immediately in PCRs or stored at −20°C. The input DNA (an aliquot of the sheared chromatin, not subjected to immunoprecipitation, but reverse-cross-linked) was diluted to 10% and used to validate the performance of different primer pairs, designed to amplify the 326-bp product. qPCR analyses were performed using IQ SYBR Green Supermix and the IQ5 Multicolor Real-Time PCR Detection System, according to the manufacturer's protocols. The primer sequences used to amplify the GH promoter region that includes the GH1 and GH2 binding sites were as follows: P1S, sense 5′-GCGGTGGAAAGGTAAGATCA-3′; and P1AS, antisense, 5′-TGCATGCCCTTTTTATACCC-3′. In addition, qPCR analysis was performed using a primer set located in a region upstream of our targeted GH promoter region. The primers upstream included the following: C1S, sense 5′-GATCTCCAACCCCCTCTGAT-3′; and C1AS, antisense, 5′-AGGGGGAGAAATTGACACCT-3′. Fold enrichment relative to the IgG IP samples was calculated using the slope of the standard.

Generation of sERα-KO mouse model

Cre recombinase downstream from the GH promoter (GHp) transgenic mice (GHp-Cre⁺) were created on a C57BL/6J and FVB/NHsd background (22). ERα floxed (ERα^fl/fl) mice from a C57BL/6J, SJL, and CD1 background, where exon 3 of the ERα (ESR1) was flanked by 2 loxP sites, were produced as described previously (23). Cre recombinase expression in somatotrophs was predicted to result in cell-selective excision of the floxed exon of ERα in somatotroph cells. The GHp-Cre⁺/ERα^fl/fl conditional knockout mice, designated as sERα-KO, were generated by first mating a female GHp-Cre⁺ mouse to a male ERα^fl/fl mouse and then back-crossing a heterozygous GHp-Cre⁺/ERα^fl/wt female from F1 with the GHp-Cre⁻/ERα^fl/fl male parent. GHp-Cre⁺/ERα^wt/wt, GHp-Cre⁻/ERα^fl/wt, and GHp-Cre⁻/ERα^fl/fl were used as controls for all experiments. No differences were observed among the control phenotypes.

Genotyping and DNA extraction

Tail biopsy samples were digested overnight at 55°C in buffer containing 10 mM Tris (pH 8), 100 mM NaCl, 10 mM EDTA, 0.1 mg/mL proteinase K, and 0.5% SDS. Samples were then incubated for 1 hour at 37°C with RNase, and genomic DNA was further isolated by performing DNA phenol-chloroform extraction and precipitated by 2-propanol. For all samples, DNA concentration was measured by spectrophotometry using an Epoch Multi-Volume Spectrophotometer System, and PCR was then performed using 100 ng of tail DNA extracted per reaction. Primers used were as follows: Cre, sense 5′-CGA CCAAGTGACAGCAATGCT-3′ and antisense 5′-GGTGCT AACCAGCGTTTTCGT-3′ (24) and mERα loxP, sense 5′-AAGGCTGCAAGGCTTTCTTTAAG-3′ and antisense 5-CCAAGGAGAACAGAGAGACTTACTAG-3′ to detect ERα floxed alleles; and knockout, sense 5′-AGGCTTTGTCTCGCTTTCC-3′ and antisense 5′-CCAAGAACAGAGAGACTTACTAG-3′ to detect knockout alleles. PCR protocols were as follows: GHp, 94°C for 2 minutes, 30 cycles at 94°C for 30 seconds, 55°C for 30 and 72°C for 40 seconds, and last cycle at 72°C for 7 minutes; and ERα loxP, 94°C for 2 minutes, 35 cycles at 94°C for 15 seconds, 60°C for 30 seconds, and 72°C for 30 seconds, and last cycle at 72°C for 7 minutes.

For identifying tissue-specific ERα floxed expression, different tissues (pituitary, hypothalamus, cerebellum, brain cortex, testis, liver, lung, heart, skeletal muscle, and kidney) from sERα-KO mice were collected. DNA from these samples was obtained by phenol-chloroform extraction and 2-propanol precipitation.

In vivo treatments and serum hormone level measurements

For our in vivo experiment, we used modified protocols by Frasor et al (25) and Lee et al (26). Two- to 3-month-old mice were anesthetized with ketamine/xylazine and ovariectomized (OVX) via a dorsal incision, followed by 11 to 14 days of recovery. Mice were injected for 3 days (1 injection/d sc) with vehicle, E₂ (40 μg/kg/d), PPT (1 mg/kg/d), or DPN (1 mg/kg/d), dissolved in DMSO and further diluted 3 times in sesame oil. At 24 hours after the final injection, in the morning, the mice were killed by deep anesthesia with isoflurane (Vedco, Inc) and cervical dislocation. The blood of each animal was collected by heart puncture, and sera were separated by centrifugation at 4000 × g for 15 minutes at 4°C and stored at −80°C until they were measured. Pituitary glands and hypothalami were obtained, immediately placed in liquid nitrogen, and then stored at −80°C for mRNA extraction.

Serum GH hormone level measurements were performed using Milliplex Rat Pituitary Immunoassay kits (Millipore Corporation) and measured by a Luminex 200 System (Luminex). PRL levels were measured using Mouse/Rat Prolactin ELISA kits (Calbiotech).

Estrous cycle assessment

Vaginal cytology samples of 2- to 3-month-old control and sERα-KO mice were prepared daily at 10:00 am over a period of 12 days. Slides were stained with a Diff-Quick stain kit (IMEB) and examined according to the protocol of Nelson et al (27) to determine the estrous cycle.

Immunofluorescence staining

Animals were anesthetized with ketamine/xylazine and perfused using 4% cold paraformaldehyde and after removal were transferred to 10% formalin. Pituitaries were washed twice with 1× PBS and transferred to 70% ethanol overnight at 4°C. The pituitaries were processed for paraffin embedding, and 6-μm sections were cut. Pituitary sections were deparaffinized and serially rehydrated in ethanol, followed by an antigen retrieval step with sodium citrate at 98°C for 20 minutes. Once sections were blocked with 3% horse serum/1× PBS solution, somatotrophs were labeled using a guinea pig anti-rat GH primary antibody (1:500 dilution; National Institute of Diabetes and Digestive and Kidney Diseases National Hormone and Peptide Program), and ERα receptors were labeled with rabbit primary antibody (1:500 dilution, ERα, H-184, sc-7207; Santa Cruz Biotechnology). After overnight incubation at 4°C, cyanine-3 (Cy3)– and fluorescein isothiocyanate (FITC)–conjugated secondary antibodies were applied to the sections followed by a 1-hour incubation at room temperature. Sections were then washed with 1× PBS and mounted with Vectashield mounting medium with 4′,6′-diamidino-2-phenylindole. Slides were viewed using a fluorescence inverted microscope (Axioskop 2; Zeiss) equipped with a charge-coupled device digital camera for image capture and processing with Axiovision (Zeiss) software. Images were captured using appropriate fluorescence filters for 4′,6′-diamidino-2-phenylindole, FITC, and Cy3 detection. Overlay images of different fluorescence signals were generated using Axiovision software. Magnification shown is ×400 using a ×40 oil immersion lens.

Statistical analyses

Statistical analyses were performed using Student unpaired t tests. Values are presented as means ± SEM. Multiple comparisons were analyzed by two-way ANOVA followed by a Bonferroni post hoc test.

Results

ERα and ERβ are expressed in both GH₃ and MtT/S cells

RT-PCR and Western blot analysis were used to confirm the presence of ERα and ERβ in both GH₃ and MtT/S cell lines. PCR bands corresponding to ERα and ERβ products were compared with those in various rat tissues, confirming the presence of the appropriate sized band in the rat cell lines and the rat pituitary (Figure 1A–C). Western blot analysis was also performed to measure ERα and ERβ protein expression in GH₃ and MtT/S cell lines before and after 10⁻⁸ M E₂ treatment. These experiments confirmed the presence of ERα and ERβ proteins in both cell lines and no change in the expression of either receptor after E₂ treatment by Western blot analysis (Figure 1D).

Figure 1.

ERα and ERβ gene expression. A–C, Representative 1% agarose gel of RT-PCR demonstrating ERα (241 bp) and ERβ (258 bp) mRNA expression in GH₃ and MtT/S cells (A) and tissue panel from rat (B and C). D, ERα and ERβ protein expression in GH₃ and MtT/S cells. Representative immunoblots for ERα (66 kDa), ERβ (55 kDa), and β-actin (42 kDa) proteins in control (C) or 10⁻⁸ M E₂-treated cells.

E₂ treatment increased GH mRNA expression in GH₃ and MtT/S cells

Time course and concentration-response experiments were performed to determine the effect of E₂ on GH mRNA expression in GH₃ and MtT/S cells. qRT-PCR analysis was performed, and the relative fold differences in mRNA levels in E₂ and vehicle-treated control cells were calculated. The results from the time course experiments showed that E₂ significantly increased GH mRNA levels after 18 hours of incubation, and these levels were sustained up to 24 hours after treatment (Figure 2A). These experiments were then repeated in GH₃ and MtT/S cells using different concentrations of E₂ (10⁻¹⁰–10⁻⁶ M) and 18 hours of incubation (Figure 2B). In both GH₃ and MtT/S cells, concentrations of 10⁻¹⁰ to 10⁻⁷ M E₂ significantly increased relative GH mRNA expression vs that in control cells treated with vehicle. At a concentration of 10⁻⁸ M, E₂ maximally stimulated GH mRNA production in GH₃ cells, whereas 10⁻⁹ M E₂ maximally stimulated GH mRNA in MtT/S cells (Figure 2B).

Figure 2.

In vitro effect of E₂ and the ER antagonist ICI 182,780 (ICI) on E₂-induced stimulation of GH mRNA expression. A, Time course experiments. Cells were incubated with E₂ (10⁻¹⁰ M) for different periods of time (6, 18, 24, or 48 hours) (n = 6). B, Concentration-response experiments. Cells were stimulated for 18 hours with increasing concentrations of E₂ (10⁻¹⁰–10⁻⁶ M) (n = 12). C and D, GH₃ and MtT/S cells were incubated for 18 hours in the absence or presence of E₂ and/or ICI 182,780 (10⁻⁸ M). GH mRNA levels of E₂-treated cells (ICI 182,780 costimulated or not) were compared with the levels of non-E₂–treated control cells. GH mRNA levels were measured by qRT-PCR. Significant differences compared with control groups: *, P ≤ .05; **, P ≤ .01; ***, P ≤ .001. Data are means ± SE of relative mRNA and are graphed as fold increases in levels relative to those of the vehicle-stimulated control (C) cells.

ICI 182,780 abolished the E₂-induced increase in GH mRNA expression in GH₃ and MtT/S cells

To confirm that the ER mediated the E₂-induced increase in GH gene expression, the ER antagonist ICI 182,780 was used. GH₃ and MtT/S cells were incubated either with E₂ 10⁻⁸ and 10⁻⁹ M, respectively, ICI 182,780 (equivalent concentration), or both. Corresponding controls (vehicle or ICI 182,780 alone) were used for comparisons. GH mRNA levels were measured by qRT-PCR for both treatment groups (ICI 182,780 and non-ICI 182,780–treated cells) and the relative fold difference of the E₂-stimulated cells was calculated. In the presence of ICI 182,780, E₂-induced increases in GH mRNA expression were completely abolished in both GH₃ and MtT/S cell lines (Figure 2, C and D).

GH and Prl are differentially regulated by ERα and ERβ in GH₃ and MtT/S cells

To distinguish further between the effects of ER isoforms, we conducted experiments with specific ERα and ERβ agonists, PPT and DPN, respectively. Concentration-response experiments were performed to determine the effect of PPT and DPN on GH mRNA expression in GH₃ and MtT/S cells (Supplemental Figure 1 published on The Endocrine Society's Journals Online web site at http://mend.endojournals.org.). GH₃ and MtT/S cells were treated with E₂ (10⁻⁹ M)_, PPT (10⁻⁹M), or DPN (10⁻¹⁰M) for 18 hours and the relative expression of GH and PRL mRNA levels was measured by qRT-PCR. E₂ treatment leads to significant increases in both GH and PRL in GH₃ cells as well as in MtT/S cells (Figure 3, A and B). Similar significant increases in GH and PRL expression were also seen after treatment with PPT in GH₃ and MtT/S cells (Figure 3, A and B). DPN treatment of GH₃ and MtT/S cells resulted in significantly increased GH mRNA levels; however, no increase was noted in PRL expression in either cell line (Figure 3, A and B). These experiments were repeated to determine the effect of E₂ or the specific receptor agonists on GH and PRL secretion as well as protein expression. GH₃ cells demonstrated a significant increase in GH secretion after a 24-hour incubation period with E₂, DPN, and PPT compared with that for the control group (Figure 3C). No significant differences in GH protein expression were noted for any of the treatment groups during the 18- and 24-hour incubation periods (Figure 3D). Although transcription increased, protein levels remain unchanged in these studies because there is probably a balanced increase in GH secretion. PRL secretion was significantly greater after 18 hours of treatment in both the E₂ and PPT groups compared with that for controls. In addition, 24 hours of treatment with E₂ and PPT resulted in significantly higher levels of PRL secretion compared with those for controls. In contrast, no differences in PRL protein expression were noted in the DPN group at either treatment time point (Figure 3F).

Figure 3.

In vitro effect of E₂, PPT, and DPN on GH and PRL expression. A and B, GH₃ (A) and MtT/S (B) cells were incubated for 18 hours in the presence or absence of E₂, PPT, or DPN (n = 6–8). GH and PRL mRNA levels were measured by qRT-PCR. C and D, GH secretion and protein expression in GH₃ cells incubated in the presence or absence of E₂, PPT, or DPN for 18 and 24 hours (n = 3–6). E and F, PRL secretion and protein expression in GH₃ cells incubated in the presence or absence of E₂, PPT, or DPN (n = 3–6). Significant differences compared to control groups: *, P ≤ .05; **, P ≤ .01; ***, P ≤ .001. Data are means ± SE of relative mRNA and are graphed as fold increases in levels relative to those of vehicle-stimulated control (C) cells.

E₂, PPT, and DPN increased POU1F1 mRNA expression in GH₃ and MtT/S cells

To understand the cellular events that may mediate GH expression, the role of the transcription factor POU1F1 was studied. GH₃ and MtT/S cells were treated for 18 hours with E₂ at increasing concentrations from 10⁻¹⁰ to 10⁻⁶ M, and POU1F1 mRNA expression was measured by qRT-PCR. In both cell lines, E₂-stimulated POU1F1 mRNA expression was significantly increased at concentrations of 10⁻⁹ and 10⁻⁸ M (Figure 4A). The effect of E₂ on POU1F1 mRNA levels in GH₃ cells was maximal at a concentration of 10⁻⁸ M and in MtT/S cells at a concentration of 10⁻⁹ M (Figure 4A). GH₃ and MtT/S cells were also treated for 18 hours with PPT and DPN. POU1F1 mRNA levels were then measured by qRT-PCR. PPT and DPN significantly stimulated POU1F1 mRNA levels similarly in GH₃ cells (Figure 4B). In MtT/S cells, significant increases were also seen after PPT and DPN treatments (Figure 4B).

Figure 4.

In vitro effect of E₂, PPT, and DPN on POU1F1. A, POU1F1 mRNA levels were measured in GH₃ and/or MtT/S cells. Cells were incubated for 18 hours with increasing concentrations E₂ (10⁻¹⁰–10⁻⁶ M). mRNA was analyzed by qRT-PCR, and the relative fold difference was calculated (n = 6). B, POU1F1 mRNA levels were analyzed in GH₃ and MtT/S cells after 18 hours of treatment with E₂ (10⁻⁸ or 10⁻⁹ M), PPT (10⁻⁹ M), or DPN (10⁻¹⁰ M). POU1F1 mRNA levels were measured by qRT-PCR and the relative fold difference was calculated (n = 4–5). C, MtT/S cells were incubated for 18 or 24 hours with E₂ (10⁻⁹ M), PPT (10⁻⁹ M), or DPN (10⁻¹⁰ M). POU1F1 protein expression was analyzed by Western blotting. Total levels of β-actin were not altered by the treatments. Relative densitometric values of protein were quantified, and means are graphed as fold increases relative to values for controls (n = 3–4). D, Effect of E₂, PPT, and DPN on POU1F1 binding to the GH promoter in MtT/S cells. GH₃ cells were incubated for 18 hours in the presence of E₂ (10⁻⁸ M), PPT (10⁻⁹ M), or DPN (10⁻¹⁰ M). ChIP analysis followed by qPCR was performed. Data were quantified and calculated as fold differences between the IgG IP DNA and the treated groups: E₂, PPT, and DPN (n = 3–4). Significant differences compared with control (C) groups: *, P ≤ .05; **, P ≤ .01; ***, P ≤ .001. Data are means ±SE. C, control.

E₂, PPT, and DPN increased POU1F1 protein expression in MtT/S cells

Western blot analysis using an antibody specific for POU1F1 was performed using protein from MtT/S cells incubated for 18 or 24 hours with E₂ (10⁻⁹ M), PPT (10⁻⁹ M), or DPN (10⁻¹⁰ M) and quantified by densitometry (Figure 4C). Increased POU1F1 protein levels were noted in all treated groups at both time points. After 18 hours of treatment, E₂ increased POU1F1 protein levels compared with those for the control group. In addition, similar increased POU1F1 protein levels were also observed in the 24-hour treatment group (Figure 4C).

E₂, PPT, and DPN increased POU1F1 binding to the GH promoter in MtT/S cells

MtT/S cells were treated for 18 hours with E₂ (10⁻⁹ M), PPT (10⁻⁹ M), or DPN (10⁻¹⁰ M) to determine whether E₂ or the selective agonists of ERα and ERβ, respectively, induced POU1F1 binding to the GH promoter. qPCR analysis of ChIP DNA was performed using primers designed at the POU1F1 binding domain of the proximal GH promoter and a primer set located in a region upstream of our targeted GH promoter designed as a control region. qPCR analysis revealed that in the presence of E₂, PPT, and DPN, POU1F1 binding to the GH promoter was increased compared with that for vehicle-stimulated cells (Figure 4D).

Generation of sERα-KO mouse model

To determine the in vivo effect of E₂ on GH expression, a genetically modified mouse model bearing a deletion of ERα in somatotrophs was constructed. Female mice expressing a transgene containing Cre recombinase downstream from the GH promoter (rGHpCre) were crossed with male mice containing loxP sites flanking exon 3 of the ERα (ESR1) (ESR1 flox/flox mouse) (23). This promoter fragment was previously shown to confer high levels of hormonally regulated expression to the somatotroph (22). To identify the presence of Cre recombinase and ERα floxed alleles, PCR genotyping of extracted genomic tail DNA was performed (Figure 5A). Furthermore, a PCR genotype analysis for ERα alleles was performed in different tissues harvested from sERα-KO mice (Figure 5B). The combined floxed allele was only detected in the pituitary, whereas in other tissues the floxed allele remained intact. In addition, paraffin-embedded slices from pituitaries of control and sERα-KO mice were prepared, and double immunofluorescence staining using GH and ERα primary antibodies was performed (Figure 5C). ERα-positive cells were visualized using Cy3 (red) second antibody, and GH-positive cells were visualized using FITC (green) secondary antibody. Double immunostaining demonstrates a lack of ERα and GH costaining in sERα-KO pituitaries (Figure 5C).

Figure 5.

Generation of sERα-KO mouse model. A, Tail PCR genotyping. A representative PCR analysis gel demonstrating GHp Cre (300 bp), homozygote wild-type (171 bp), heterozygote (both 171 and 259 bp), and homozygote (259 bp) floxed ERα bands. B, A representative tissue panel from a sERα-KO mouse demonstrating the presence of homozygous floxed alleles (882 bp) in all tissues and the knockout product after Cre recombination in the pituitary only (224 bp). C, double immunofluorescence staining for GH and ERα of control and sERα-KO mouse pituitaries. Paraffin-embedded pituitary slices from control and sERα-KO mice were double stained against GH (red) and ERα (green) proteins. There is a lack of GH and ERα coimmunostaining in sERα-KO mice.

Body weight and length of both male and female sERα-KO mice as well as those of mice used as controls were measured 3 times per week. Data indicate that sERα-KO mice were of similar size and had growth rates similar to those of their littermate controls (Supplemental Figure 2, A and B). There were no differences between sERα-KO and control mice in the timing of puberty or the growth rate during puberty. sERα-KO mice had normal estrous cycles, were fertile, and lactated normally, and there was no difference in the number of pups delivered compared with that for the control mice (data not shown).

Differences in pituitary mRNA levels between control and sERα-KO mice

RNA from pituitaries of 4- to 6-month-old control mice and sERα-KO mice was extracted and analyzed for GH, PRL, POU1F1, ERα, and ERβ genes using qRT-PCR. Data were calculated by the ΔΔC_t method and are presented as fold difference (females and males separately). Results indicate that in both male and female mice, the relative expression of GH, PRL, POU1F1, and ERα was significantly lower in sERα-KO mice than in control sex-matched littermates (Figure 6, A–D). In contrast, no differences in ERβ mRNA expression levels were seen in control and sERα-KO mice of either sex (Figure 6E).

Figure 6.

Basal gene expression studies in control and sERα-KO mice. Basal gene expression of control and sERα-KO mice, males and females, respectively, was compared by using qRT-PCR analysis. Significant differences compared with control (C) groups: *, P ≤ .05 (n = 7–14).

Stimulation by E₂ demonstrates receptor-selective increases in both GH and PRL gene expression in the sERα-KO mouse model

To monitor the efficacy of E₂, PPT, and DPN treatment in vivo, uteri from experimental mice were weighed. Supplemental Figure 3A represents the average uterine weights, calculated as a percentage of the whole-body weight for each animal in each treatment group. Three weeks after ovariectomy, the average uterine weights were 0.16% and 0.19% of the body weights in control and sERα-KO mice, respectively. Treatment with E₂ resulted in significantly increased uterine weights in control and sERα-KO mice. PPT-treated animals had also uterine weights significantly higher than those of the control animals, whereas DPN-treated animals had uterine weights equivalent to those of vehicle-treated OVX females (Supplemental Figure 3A). The uterine weight data from control, E₂-, PPT-, or DPN-treated animals were compared with the uterine weights of age-matched non-OVX females from different stages of the estrous cycle (Supplemental Figure 3B). In the non-OVX animals, the uterus weight changed during the estrous cycle, and there was no difference in the uterus weight between control and sERα-KO mice at each stage of the estrous cycle.

Ovariectomies were performed on both sERα-KO and control female mice, which were then treated with vehicle, E₂, DPN, or PPT. Pituitary glands were dissected from mice, and mRNA was extracted to measure relative GH and PRL mRNA expression levels using qRT-PCR analysis. In control animals, E₂ treatment resulted in an increase in GH mRNA levels compared with those in vehicle-treated mice (Figure 7A). Significant increases in GH mRNA were also seen in PPT- and DPN-treated mice. In sERα-KO mice, similar to the control mice, E₂ and DPN significantly increased pituitary GH mRNA expression levels, but PPT had no significant effect on GH mRNA expression levels. PRL mRNA expression levels were also significantly higher in control mice after E₂ and PPT treatment (Figure 7B). No significant differences in PRL mRNA expression levels were seen in DPN-treated control mice. The PRL mRNA expression levels in sERα-KO mice increased in E₂- and PPT-treated mice (Figure 7B). In sERα-KO mice, DPN treatment also had no significant effect on PRL mRNA levels (Figure 7B).

Figure 7.

Effect of E₂, PPT, or DPN on GH and PRL mRNA and serum GH and PRL levels in control and sERα-KO mice. Control and sERα-KO female mice were OVX and then treated with vehicle (control [C]), E₂, PPT, or DPN. Whole pituitary RNA was extracted and qRT-PCR analysis for GH (A) and PRL mRNAs (B) was performed. Data are fold differences compared with controls (n = 3–14). Serum levels (nanograms per milliliter) of GH (C) and PRL concentrations (D) were measured from control, E₂-, PPT-, or DPN-treated OVX mice by Luminex or ELISA, respectively (n = 6–28). Significant differences compared with control groups: *, P ≤ .05; **, P ≤ .01; ***, P ≤ .001.

Receptor-selective changes in serum GH and PRL levels in the sERα-KO mouse model

Serum was obtained from OVX sERα-KO and control female mice treated with vehicle, E₂, DPN, and PPT. GH and PRL hormone levels were measured by Luminex and ELISA assays, respectively. In the control OVX mice, the average serum GH level in the vehicle-treated animals was 0.41 ng/mL (Figure 7C). Serum GH levels were significantly increased in all treatment groups after E_2, PPT, or DPN treatments. The average serum PRL levels in the OVX mice increased with E₂ and PPT treatment and were unchanged in the DPN treatment group (Figure 7D).

In OVX sERα-KO mice, serum GH and PRL levels increased after E₂ treatment (Figure 7, C and D). PPT significantly stimulated PRL secretion but not serum GH levels. DPN stimulated GH secretion but did not significantly increase serum PRL levels (Figure 7, C and D).

Discussion

GH levels have been shown to vary during the stages of mammalian development. GH levels increase during puberty, peak in late puberty, and decrease with aging (21, 28). In addition, there is sexual dimorphism in the GH secretion profile (29–32) with women typically having an overall higher 24-hour integrated GH concentration (33, 34) and higher circulating GH concentrations (35–37). In women, the GH secretion profile closely correlates with ovarian function, such that during the menstrual cycle, changes in serum GH levels directly correlate with changes in estrogen levels (1, 38–40). In sheep, treatment with estrogen results in increased baseline plasma GH concentrations (41), and a concomitant surge in GH occurs at the time of the spontaneous or E₂-induced LH surge (42, 43). In OVX or intact baboons and macaques, E₂ provokes an increase in serum GH and IGF-I (44, 45). In mice, low levels of GHRH receptor, POU1F1, and GH expression were found in aromatase KO females, suggesting a role for estrogen in regulating their expression (46). Pregnancy and menopause are also associated with changes in the GH profile (47). In humans, as well as in rodents, GH levels increase during the second half of pregnancy (47–49) and closely correlate with levels of E₂ (50). Conversely, the dramatic decrease in estrogens after menopause is accompanied by a decrease in serum GH concentrations (33, 34). Further, oral estrogen administration in postmenopausal women increases average 24-hour serum GH concentrations, GH pulse amplitude, and basal GH concentrations (51). Mean plasma GH levels are higher in premenopausal women than in postmenopausal women or in young men (1). Women taking oral contraceptives have higher plasma concentrations of GH (52). Furthermore, estrogen treatment of healthy male subjects results in enhanced basal GH secretion (53), and Parker et al (54) reported that gonadal steroid administration elevated serum IGF-I levels in young GH-sufficient males, which was opposite to the observation in GH-deficient patients.

Despite abundant evidence regarding GH responsiveness to sex steroid modification in mammals, the neuroendocrine targets of E2 regulating GH expression and secretion have not been clearly defined. A direct effect of estrogen on somatotrophs was demonstrated by Jansson et al (55) with continuous GH secretion from pituitary autotransplants to the kidney capsule stimulated by estrogens. The reports of estrogen effects on GH release from pituitary cells in vitro are conflicting (2). It has been reported that estrogens, when added to pituitary cell culture systems may lead to an increase, decrease, or no change in GH secretion into the media (1). These conflicting results could be due to the different models and conditions used for each experiment and the role of other factors not accounted for in these experiments.

In the present study, 2 different GH-secreting cell lines, GH₃ and MtT/S, were used to determine the effects of E₂ on GH gene expression and secretion. Both lines are derived from rat MtT tumors and have somatotroph characteristics. GH₃ cells have been described as somatolactotroph cells, whereas MtT/S have been described as pure somatotrophs (56). ERα, which is localized in most pituitary cells (57, 58), is thought to be expressed in a higher percentage of pituitary cells than ERβ (59) and thought to mediate the direct effects of estrogen in the pituitary. Both GH₃ and MtT/S cells express ERs with ERα isoform expression mRNA levels higher than ERβ mRNA levels (7).

Our results confirm previous observations that these cell lines express mRNA and protein for both ERα and ERβ (7, 60). We show that the stimulatory effect of E₂ on GH gene expression in both cell lines is time and concentration dependent, which correlates with previous studies that demonstrated the direct effect of estrogens on GH synthesis. The ER-selective antagonist ICI 178,780 completely abolishing the observed stimulatory effect of E₂ in both GH₃ and MtT/S cells confirms that activation of the classic nuclear hormone receptor pathway regulates gene expression. To determine the ER isoform or isoforms that mediate this effect, synthetic ERα- or ERβ-selective agonists, PPT and DPN, respectively, were used in concentration-response experiments. It has been demonstrated that PPT is a potent ERα agonist, with a 400-fold preference for ERα over ERβ (61, 62). In contrast, DPN is a selective agonist for ERβ with more than 70-fold higher binding activity than ERα (25, 63). We determined that both PPT and DPN increased GH mRNA expression and secretion in GH₃ and MtT/S cells. E₂ also stimulated PRL mRNA expression and secretion in vitro. We show that this is an ERα-mediated event because PPT increased PRL mRNA, whereas DPN did not significantly affect PRL mRNA expression. E₂, PPT, and DPN increase the expression of POU1F1, which is a major transcriptional regulator of GH gene expression and production.

Accompanying the E₂-induced increase in POU1F1 protein is an increase in binding of POU1F1 to the GH promoter at the previously defined POU1F1 binding domains, GH1 and GH2. However, the effect of E₂ on GH expression may be direct by increasing cellular levels of ERs and/or mediated by POU1F1 DNA binding. We support these in vitro data with the generation of a sERα-KO developed by crossing a transgenic female mouse expressing Cre recombinase downstream from the GH promoter to a male mouse bearing a floxed ERα gene. The presence of a recombinant ERα band specifically in the pituitary was demonstrated in a tissue panel derived from the sERα-KO mouse. Furthermore, double-label immunostaining for GH and ERα demonstrated coexpression only in pituitary somatotroph cells. The sERα-KO mice had lower levels of basal GH, PRL, POU1F1, and ERα mRNA expression in the pituitary than control littermates, whereas there was no difference in ERβ mRNA expression. Similar to the in vitro results, E₂, PPT, or DPN treatment of OVX control mice resulted in a 50% increase in pituitary GH mRNA expression and a 5- to 8-fold increase in serum GH levels. Furthermore, E₂ and PPT, but not DPN, stimulated pituitary PRL mRNA and serum levels, which is consistent with the in vitro studies. Although only 10% of lactotrophs are targeted using this experimental strategy, the contribution of the somatolactotroph to prolactin expression and secretion is not directly quantitated. In sERα-KO mice, E₂ or DPN stimulated GH synthesis and secretion, but PPT had no effect, consistent with the absence of ERα-responsive GH gene expression. The effects of E₂, PPT, or DPN on PRL gene expression and secretion were similar in the control and sERα-KO mice. After E₂ treatment, OVX sERα-KO mice had diminished levels of GH mRNA expression and secretion, and PPT had no effect in sERα-KO mice but was able to increase GH synthesis in the control mice, suggesting that the estrogen effect is mediated by ERα. Because the ERβ pathway was preserved, DPN had no differential effect on GH production in sERα-KO mice. Although E₂ has been implicated in the increase in GH associated with puberty (21, 28), sERα-KO mice were of similar size and had growth rates and puberty and pubertal growth rates similar to those of their littermate controls. Hence, sERα is not likely to have a significant role in mediating the rise in pubertal GH as demonstrated in this model.

Our in vitro and in vivo data indicate that the mechanism of action of estrogens on somatotroph hormone production involves the classic nuclear hormone receptor pathway. Our experiments demonstrate that both ER isoforms stimulate GH gene expression, whereas only ERα mediates an increase in PRL expression; the latter result confirms the generally accepted notion that the estrogen effect on PRL production is ERα dependent. It is intriguing that ER isoform specificity does not extend to GH production based on the common embryological origin of the lactotroph and somatotroph cells. Others have suggested that an important physiological role for ERβ is to regulate ERα-mediated gene transcription, although such a regulatory pathway could not explain the results observed in our study in which GH but not PRL mRNA levels were regulated by DPN, an ERβ ligand (64). The results from our experiments using GH₃ and MtT/S cells show that E₂, either through ERα or ERβ activation, stimulates POU1F1 synthesis and binding to the GH1/GH2 sites of the proximal GH promoter. In addition, the observed lower basal POU1F1 mRNA levels seen in sERα-KO mice compared with those in control mice suggest that ERα signaling regulates GH production at least in part through changes in POU1F1 expression.

In summary, the results of the present in vitro and in vivo studies demonstrate that estrogens can directly stimulate GH synthesis and secretion. In the somatotroph, these effects are mediated by the classic nuclear hormone receptor pathway, the homeodomain protein, POU1F1, and involve both ERα and ERβ isoforms. These results are in contrast to those found in the lactotroph where only ERα stimulates Prl expression and demonstrate, for the first time, a consistent differential effect of ER isoforms on pituitary gene expression of Prl and Gh in vitro and in vivo.

Additional material

Supplementary data supplied by authors.