Abstract
The G protein-coupled receptor Gpr30 (Gper) was recently claimed to bind to estradiol and to activate cytoplasmic signal transduction pathways in response to estradiol. However, there are conflicting data regarding the role of Gpr30 as an estrogen receptor (ER): several laboratories were unable to demonstrate estradiol binding to GPR30 or estradiol-activated signal transduction in Gpr30-expressing cells. To clarify the potential role of Gpr30 as an ER, we generated Gpr30-deficient mice. Although Gpr30 was expressed in all reproductive organs, histopathological analysis did not reveal any abnormalities in these organs in Gpr30-deficient mice. Mutant male and female mice were as fertile as their wild-type littermates, indicating normal function of the hypothalamic-pituitary-gonadal axis. Moreover, we analyzed estrogenic responses in two major estradiol target organs, the uterus and the mammary gland. For that purpose, we examined different readout paradigms such as morphological measures, cellular proliferation, and target gene expression. Our data demonstrate that in vivo Gpr30 is dispensable for the mediation of estradiol effects in reproductive organs. These results are in clear contrast to the phenotype of mice lacking the classic ER alpha (Esr1) or aromatase (Cyp19a1). We conclude that the perception of Gpr30 (based on homology related to peptide receptors) as an ER might be premature and has to be reconsidered.
Introduction
Estradiol has a pivotal role in reproduction, uterine physiology, and mammary gland development [1]. The analysis of mouse models in which estrogen receptor (ER) alpha (Esr1), ER beta (Esr2), or both were inactivated by gene targeting has demonstrated that mainly the classic ERs mediated the effects of estradiol in these processes [2]. However, several estrogenic responses were maintained in Esr1-deficient mice and were not blocked by the classic ER antagonist ICI182780 [3, 4]. Therefore, it was speculated that other ERs might exist. Although related to peptide receptors [5], GPR30 (G protein-coupled receptor 30) protein was claimed to bind to and signal in response to estradiol [6, 7]. It was speculated that Gpr30 might have a role in the mediation of nongenomic estradiol effects [6, 7]. In line with this hypothesis, it was demonstrated that after estradiol stimulation GPR30 activated calcium currents and cAMP signaling in transiently transfected cells [6, 7]. The ER antagonists tamoxifen and ICI182780 acted as full agonists on GPR30 [6]. In Esr1-negative, Gpr30-positive breast cancer cells, GPR30 stimulated extracellular signal-regulated kinase (ERK) phosphorylation via transactivation of the epidermal growth factor receptor (EGFR) pathway. Using a Gβγ-dependent pathway, GPR30 activated matrix metalloproteinases that released surface-bound pro-heparin-binding-EGF. Subsequently, the EGFR was activated and stimulated ERK phosphorylation [8]. The transactivation of the EGFR pathway by estrogens is an example of nongenomic estradiol signaling, which is also of importance in normal uterine and mammary gland biology. Several estrogenic responses in the uterus (i.e., uterine epithelial cell proliferation) and in the mammary gland (i.e., ductal elongation and end bud growth) are mediated via estradiol-activated EGFR transactivation and can be blocked by the administration of neutralizing EGF antibodies [9, 10]. Although the classic Esr1 was shown to transactivate the EGFR pathway [11], it was speculated that GPR30 might also be involved in mediating estrogenic responses in reproductive organs via transactivation of this pathway [12]. A possible role for GPR30 in the regulation of reproductive functions was further supported by the finding that Gpr30 mRNA expression in the uterine and mammary gland epithelium was regulated in an estrous cycle-dependent manner [13]. In addition, gonadotrophic hormones influenced the expression of GPR30 protein in granulosa and theca cells of hamster ovaries. Because of these findings, it was suggested that Gpr30 might be involved in the regulation of preantral follicle development [14].
To define the physiological role of GPR30 in reproductive biology, we generated Gpr30-deficient mice. We analyzed reproductive function and typical estradiol responses, including those mediated via EGFR transactivation. The aim of this study was to clarify whether the presumptive ER GPR30 had any effect on the mediation of estrogenic responses in classic target organs such as the uterus and the mammary gland.
Materials and Methods
Chemicals
Chemicals used included 17β-estradiol and 5′-bromo-2′-deoxyuridine (BrdU). These were purchased from Sigma.
Generation of Mice
To inactivate Gpr30 in vivo, we deleted exon 3 from the murine genome, which encodes the complete open reading frame of the GPR30 protein. The targeting construct (Fig.1a, second panel) was based on a 9.8-kB genomic fragment representing the genomic sequence of exons 1, 2, and 3 and the surrounding introns of the murine Gpr30 gene (Fig. 1a, top panel). This fragment, obtained from an RP23 BAC library, was modified using homologous recombination in Escherichia coli [15] to carry a loxP site in the 5′ direction of exon 3, a PGKtkneo cassette flanked by frt sites, and a loxP site in the 3′ direction of exon 3. C57BL/6N embryonic stem cells were transfected with the linearized targeting construct. After transfection of the embryonic stem cells, G418-resistant clones were analyzed by Southern blot using probes from outside of the homology arms of the targeting vector (depicted as A and B in Fig. 1a). The 3′-external probe B (Fig. 1a, top panel) detected a 16.9-kB NheI fragment from the wild-type allele in all clones and an additional 5.7-kB NheI fragment in all clones that were homologously recombined (Fig. 1b). The 5′-external probe A (Fig. 1a, top panel) detected a 10-kB KpnI fragment from the endogenous allele in all clones and an additional 12-kB KpnI fragment in one allele that underwent homologous recombination with the targeting vector (Fig. 1c). The internal probe C (Fig. 1a, second panel), derived from the PGKtkneo cassette, detected an 8.3-kB HpaI fragment in all clones that were homologously recombined and that had a single integration of the targeting vector (Fig. 1d). No signal was obtained from the wild-type allele. The frequency of homologous recombination was 3.2%. One homologously recombined clone (clone 7) harboring the targeted allele (Fig. 1a, third panel) was used for the generation of chimeric mice by blastocyst injection. Highly chimeric mice were bred to C57BL/6 females, and offspring heterozygous for the targeted allele (Gpertm1135Arte) were identified by Southern blot analysis. To eliminate the selection marker and exon 3, mice heterozygous for the targeted allele were bred with mice carrying one copy of the Cre recombinase transgene under the control of the ROSA26 locus (C57BL/6-Gt(ROSA)26Sortm16(Cre)Arte). The resulting offspring, heterozygous for the null allele (Fig. 1a, bottom panel), were backcrossed with C57BL/6 mice to eliminate the Cre recombinase transgene. Wild-type and mutant (Gpertm1135.2Arte) experimental animals were derived from heterozygous intercrosses and were devoid of the Cre recombinase transgene. The official nomenclature for Gpr30 has been changed to Gper (G protein-coupled ER 1). Because there is no consensus that Gpr30 acts as an ER, we will use the previous nomenclature in this article.
Generation of Gpr30-deficient mice. a) The targeting strategy that led to loss of exon 3 (black box) encoding the complete open reading frame of the GPR30 protein is depicted. A scheme of the wild-type locus, the targeting vector, and the resulting alleles is shown. Black triangles indicate loxP sites; white triangles, frt sites; K, KpnI; N, NheI; and H, HpaI. A and B represent probes outside of the homology arms used for Southern blot analysis of electroporated embryonic stem cell clones. Probe C was derived from the neomycin cassette (neo) and was used for the analysis of single-vector integration. b–d) Southern blot analysis of electroporated stem cell clones is depicted. Wt indicates a wild-type stem cell clone that was used as a control. The full integration of the 3′ arm of the targeting vector and of the distal loxP site was analyzed using an NheI digest, followed by hybridization with probe B (signal for the wild-type allele was at 16.9 kb, and signal for the targeted allele was at 5.7 kb) (b). Integration of the 5′ arm of the targeting vector was analyzed by KpnI digest, followed by hybridization with probe A (signal for the wild-type allele was at 10 kb, and signal for the targeted allele was at 12 kb) (c), and by HpaI digest, followed by hybridization with the internal probe C (no signal for the wild-type allele, and signal for the targeted allele was at 8.3 kb) (d). The HpaI digest, followed by hybridization with internal probe C, was also used to analyze single integration of the targeting vector (c). As evidenced by Southern blot analysis, clones 3, 6, and 7 represented homologously recombined embryonic stem cell clones. Clone 7 was used for the generation of chimeric mice. e) Representative results of genotyping PCR. Amplification of the wild-type allele resulted in a band of 398 bp, whereas amplification of the mutant allele resulted in a band of 560 bp. The genomic localization of the primers P1, P2, and P3 used for genotyping of the mice is shown schematically (a). f) Gpr30 expression was lost in Gpr30-deficient (−/−) mice as demonstrated by quantitative RT-PCR analysis of stomach, ovary, mammary gland, and uterus mRNA. Primers R1 and R2 used for RT-PCR analysis localized to exon 3 (a).
Mice were maintained on a 14L:10D cycle and were provided with food and water ad libitum. All animal procedures were performed according to German animal welfare law with the permission of the district government of Berlin.
Genotyping of Mice by PCR
Genomic DNA was extracted from 1- to 2-mm-long tail tips using the NucleoSpin Tissue kit (Macherey-Nagel). Genomic DNA (0.8 μl, 1:10 diluted) was analyzed by PCR in a final volume of 25 μl in the presence of 1.5 mM MgCl2, 400 μM deoxyribonucleotide triphosphates, 400 nM of each primer, and 5 U of Taq DNA polymerase (Eppendorf). The sequence of primer P1 was ACCCACAGCTCTCTTGTGTGC, the sequence of primer P2 was TCTGCGTACTCTCCTATGTACC, and the sequence of primer P3 was TCATTTTATCGCCTACTTGTTACC. The genomic localization of the genotyping primers is shown schematically in Figure 1a. Following a denaturing step at 94°C for 3 min, 35 cycles of PCR were performed, each consisting of a denaturing step at 94°C for 45 sec, followed by an annealing phase at 60°C for 1 min and an elongation step at 72°C for 1 min. The PCR was finished by a 10-min extension step at 72°C. Probes were analyzed on a 2% agarose gel. Amplification of the wild-type allele resulted in a band of 398 bp (product amplified by primers P2 and P3), whereas amplification of the mutant allele delivered a band of 560 bp (product amplified by P1 and P3) (Fig. 1e). In the case of the wild-type allele, a 3.45-kB P1/P3 PCR product apart from the P2/P3 PCR product theoretically could be generated. However, the PCR conditions (i.e., an elongation step of 1 min) did not allow for the amplification of the P1/P3 product from the wild-type allele. In the case of the mutant null allele, the binding site for primer P2 was eliminated by the targeting strategy. Therefore, only the P1/P3 PCR product resulting in a band of 560 bp was generated (Fig. 1e).
Gross Histopathological Analysis of Gpr30-Deficient Mice
Five-month-old female mutant mice (n = 9) and their wild-type littermates (n = 8) were killed. The relative organ weights were determined and are given in Table 1. Various tissues were embedded in paraffin and processed for hematoxylin-eosin staining.
Relative organ weights in female mice given as percentages of body weight.
| Organ | Wild-type | Mutant |
|---|---|---|
| Brain | 2.08 ± 0.24 | 1.96 ± 0.16 |
| Heart | 0.69 ± 0.08 | 0.68 ± 0.07 |
| Lung | 0.55 ± 0.10 | 0.55 ± 0.09 |
| Pancreas | 0.76 ± 0.11 | 0.79 ± 0.20 |
| Stomach | 0.76 ± 0.11 | 0.85 ± 0.12 |
| Liver | 5.05 ± 0.30 | 4.58 ± 0.26a |
| Kidney | 0.63 ± 0.04 | 0.65 ± 0.07 |
| Spleen | 0.35 ± 0.04 | 0.37 ± 0.08 |
| Thymus | 0.30 ± 0.06 | 0.26 ± 0.06 |
| Uterus | 0.34 ± 0.11 | 0.41 ± 0.13 |
| Organ | Wild-type | Mutant |
|---|---|---|
| Brain | 2.08 ± 0.24 | 1.96 ± 0.16 |
| Heart | 0.69 ± 0.08 | 0.68 ± 0.07 |
| Lung | 0.55 ± 0.10 | 0.55 ± 0.09 |
| Pancreas | 0.76 ± 0.11 | 0.79 ± 0.20 |
| Stomach | 0.76 ± 0.11 | 0.85 ± 0.12 |
| Liver | 5.05 ± 0.30 | 4.58 ± 0.26a |
| Kidney | 0.63 ± 0.04 | 0.65 ± 0.07 |
| Spleen | 0.35 ± 0.04 | 0.37 ± 0.08 |
| Thymus | 0.30 ± 0.06 | 0.26 ± 0.06 |
| Uterus | 0.34 ± 0.11 | 0.41 ± 0.13 |
Significantly different from wild-type (P < 0.005).
Relative organ weights in female mice given as percentages of body weight.
| Organ | Wild-type | Mutant |
|---|---|---|
| Brain | 2.08 ± 0.24 | 1.96 ± 0.16 |
| Heart | 0.69 ± 0.08 | 0.68 ± 0.07 |
| Lung | 0.55 ± 0.10 | 0.55 ± 0.09 |
| Pancreas | 0.76 ± 0.11 | 0.79 ± 0.20 |
| Stomach | 0.76 ± 0.11 | 0.85 ± 0.12 |
| Liver | 5.05 ± 0.30 | 4.58 ± 0.26a |
| Kidney | 0.63 ± 0.04 | 0.65 ± 0.07 |
| Spleen | 0.35 ± 0.04 | 0.37 ± 0.08 |
| Thymus | 0.30 ± 0.06 | 0.26 ± 0.06 |
| Uterus | 0.34 ± 0.11 | 0.41 ± 0.13 |
| Organ | Wild-type | Mutant |
|---|---|---|
| Brain | 2.08 ± 0.24 | 1.96 ± 0.16 |
| Heart | 0.69 ± 0.08 | 0.68 ± 0.07 |
| Lung | 0.55 ± 0.10 | 0.55 ± 0.09 |
| Pancreas | 0.76 ± 0.11 | 0.79 ± 0.20 |
| Stomach | 0.76 ± 0.11 | 0.85 ± 0.12 |
| Liver | 5.05 ± 0.30 | 4.58 ± 0.26a |
| Kidney | 0.63 ± 0.04 | 0.65 ± 0.07 |
| Spleen | 0.35 ± 0.04 | 0.37 ± 0.08 |
| Thymus | 0.30 ± 0.06 | 0.26 ± 0.06 |
| Uterus | 0.34 ± 0.11 | 0.41 ± 0.13 |
Significantly different from wild-type (P < 0.005).
Analysis of Reproductive Function by Continuous Mating Studies
To analyze the fertility of wild-type and Gpr30-deficient mice, we performed continuous mating studies as described previously [16]. Ten-wk-old wild-type males (n = 9) and Gpr30-deficient males (n = 9) were housed individually and were mated for 1 mo with two 8-wk-old wild-type C57BL/6 females per male. After 1 mo, the females were removed and were replaced by two new C57BL/6 females for 1 mo. The females were kept for an additional 3 wk after each 1-mo breeding period to determine the number of litters and offspring produced by each male. To study female fertility, one wild-type (n = 7) and one mutant (n = 7) female were mated with one C57BL/6 male for 4 wk. The females were then separated, and the number of litters and the litter size per female were determined. Afterward, a second mating round was performed for 1 mo using a fresh male.
Uterine Growth Assays
Gpr30-deficient females (n = 21) and their wild-type littermates (n = 28) were ovariectomized at age 10 wk. Two weeks after ovariectomy, the animals were treated for 3 days with daily subcutaneous injections of vehicle (ethanol/arachisoil 1 + 9, v/v; 14 wild types and 9 mutants) or 100 ng of estradiol (14 wild types and 12 mutants). Animals were killed on Day 4. Two hours before killing, animals were injected intraperitoneally with BrdU (70 mg/kg of body weight) dissolved in PBS. Mice were killed by cervical dislocation. The relative uterine weight was determined. One uterine horn was fixed in 4% buffered formalin and embedded in paraffin, and the other uterine horn was rapidly frozen in liquid nitrogen and processed for gene expression analysis. To evaluate uterine epithelial cell height, paraffin sections of the uterus were stained with hematoxylin-eosin. Uterine sections were examined using an Axiophot microscope equipped with an AxioCam camera (Zeiss). For each animal, five different areas were evaluated to determine the average epithelial cell height using the KS 400 Imaging System Release Program (Zeiss). To analyze uterine epithelial cell proliferation, BrdU immunostaining using a mouse monoclonal anti-BrdU antibody (M0744; DAKO) was performed as described previously [17]. The percentage of proliferating cells in the uterine epithelium was determined by counting at least 300 epithelial cells per animal.
Mammary Gland Whole-Mount Assay
Gpr30-deficient females (n = 21) and their wild-type littermates (n = 24) were ovariectomized at the beginning of 6 wk of life. Two weeks after ovariectomy, the animals were treated for 3 wk with daily subcutaneous injections of vehicle (7 wild types and 7 mutants), 100 ng of 17β-estradiol (8 wild types and 7 mutants), or 100 ng of 17β-estradiol plus 80 mg/kg of progesterone (9 wild types and 7 mutants) dissolved in ethanol/arachisoil (1 + 9, v/v). The left inguinal mammary gland was removed, spread on a glass slide, and fixed for 48 h at room temperature in Carnoy fixative (6 parts 100% ethanol, 3 parts chloroform, and 1 part glacial acetic acid). After staining in carmine alum (0.2% carmine alum, 0.5% aluminum potassium sulphate, and 1 crystal of thymol), the mammary glands were dehydrated, cleared in xylene, and stored in Protaqstura (Quartett GmbH, Berlin, Germany). For each animal, the number of end buds and side branches in six different areas distal to the lymph node was determined by an investigator who was blinded to the experimental treatment the animals had received. The dorsal two-thirds section of the right inguinal mammary gland was fixed in 4% formalin at 4°C overnight and processed for BrdU immunostaining [17]. The number of BrdU-positive ductal epithelial cells was determined by evaluating four complete transverse mammary gland sections per animal. The ventral third of the right inguinal mammary gland (without lymph node) was rapidly frozen in liquid nitrogen and processed for gene expression analysis.
Gene Expression Analysis by Quantitative RT-PCR
Uterine, mammary gland, and stomach RNA was isolated after homogenization of tissues in guanidinium thiocyanate [18]. Five micrograms of RNA was digested with deoxyribonuclease I and reversely transcribed with random hexamers using the SuperScript III First-Strand Synthesis System (Invitrogen). Real-time Taqman PCR analysis was performed using the ABI Prism 7700 Sequence Detector System according to the manufacturer's instructions (PE Applied Biosystems). Prevalidated probes and primers for murine indoleamine-pyrrole 2,3 dioxygenase (Indo) (catalog No. Mm00492586_m1), cytokeratin 18 (Krt18) (catalog No. Mm01601702_g1), lactotransferrin (Ltf) (catalog No. Mm00434787_m1), glucose-regulated protein of 78 kDa (Hspa5) (catalog No. Mm00517691_m1), Wnt4 (catalog No. Mm00437341_m1), and TATA box-binding protein (Tbp) cDNA (catalog No. Mm00446973_m1) were purchased from PE Applied Biosystems. For the analysis of Gpr30 cDNA expression, custom-made primers and probes were ordered from PE Applied Biosystems. The forward primer R1 was GTCACGCCTACCCCTTGACA, the reverse primer R2 was CCTGAAGGTCTCTCCCAGGAA, and the sequence of the probe was CCACATAGTCAACCTTG. The localization of both primers is schematically shown in Figure 1a. Relative mRNA levels were calculated using the comparative ΔCT method. In the mammary gland, Indo expression levels were normalized to Krt14 expression, whereas uterine gene expression was normalized to Tbp expression.
Ribonuclease Protection Analysis
To study the tissue distribution of Gpr30 and Esr1 mRNA, ribonuclease (RNase) protection analysis was performed as described previously [19]. Ten micrograms of RNA from various tissues obtained from adult C57BL/6 mice was subjected to hybridization. The signals obtained were quantified using a phosphoimager (Molecular Dynamics) and were normalized over those of Tbp. Probes used in RNase protection analysis were Gpr30 (nucleotides [nt] 611–950 from the murine Gpr30 mRNA, NM_029771), Esr1 (nt 713–910 from the murine Esr1 mRNA, NM_007956), and Tbp (nt 117–258 from the murine Tbp mRNA, NM_013684).
Statistical Analysis
Data are given as mean ± SD. Data were analyzed using two-way ANOVA with genotype and treatment as fixed effects. Because heterogeneity between the groups could not be excluded, a model with different residual variance for each combination of genotype and treatment was fitted. The analysis was performed using PROC MIXED in SAS, version 9.1 (SAS Institute).
Results
To determine the tissue distribution of Gpr30 and Esr1, we performed RNase protection analysis using various adult mouse tissues (Fig. 2). Highest expression of Gpr30 was found in the stomach, followed by the adrenal gland and the spinal cord. In addition, Gpr30 was found to be expressed in the following organs: several brain regions, duodenum, colon, spleen, lung, kidney, heart, muscle, ovary, uterus, and mammary gland. No Gpr30 expression was detectable in the liver, thymus, or testis. In general, Esr1 was much more abundant than Gpr30 in the liver and the reproductive organs. Both receptors were expressed to a similar extent in the central nervous system. Within the hypothalamus, Esr1 was expressed to greater extent than Gpr30.
Tissue distribution of Gpr30 and Esr1 as evidenced by RNase protection analysis. Relative mRNA levels (normalized to Tbp expression) of Esr1 (black bars) and Gpr30 (gray bars) are depicted for the following organs: cerebral cortex (1), hippocampus (2), hypothalamus (3), cerebellum (4), spinal cord (5), stomach (6), duodenum (7), colon (8), liver (9), thymus (10), spleen (11), lung (12), kidney (13), heart (14), skeletal muscle (15), ovary (16), uterus (17), mammary gland (18), testis (19), and adrenal gland (20).
To assess the physiological role of Gpr30, we deleted this gene from the murine genome. Exon 3 of the murine Gpr30 locus encoded the complete translated sequence of the receptor protein, as well as parts of the 3′- and 5′-untranslated regions [5]. To completely inactivate Gpr30 in vivo, we deleted exon 3 as described in Materials and Methods and as shown in Figure 1a. Homologous recombination of embryonic stem cell clones was analyzed by Southern blot analysis (Fig. 1, b–d) as described in Materials and Methods and in the figure legend. According to our targeting strategy, the neomycin selection cassette was removed from the murine genome (Fig. 1a). The experimental wild-type and mutant animals were littermates and were obtained from heterozygous intercrosses. All mice were on a pure C57BL/6 genetic background. A typical genotyping PCR is shown in Figure 1e. The PCR genotyping strategy is described in detail in Materials and Methods. As evidenced by quantitative RT-PCR, Gpr30 mRNA expression was completely absent in Gpr30-deficient mice (Fig. 1f). As examples, we analyzed the expression of Gpr30 mRNA in the stomach, ovary, mammary gland, and uterus of wild-type and Gpr30-deficient mice. Quantitative RT-PCR (Fig. 1f) and RNase protection analysis (Fig. 2) each independently revealed that Gpr30 expression was approximately 10 times higher in the stomach compared with the expression in reproductive organs (compare relative Gpr30 expression in Fig. 2 and Fig. 1f in the respective organs). At the age of weaning, Gpr30-deficient mice were found at the expected mendelian ratio. Female mutants appeared healthy and were indistinguishable from their wild-type littermates. At the age of 5 mo, the body weight of wild-type females was 25.3 ± 0.7 g, and the body weight of mutant females was 25.3 ± 0.9 g. Relative organ weights (expressed as percentage of body weight) of 5-mo-old wild-type and mutant females are given in Table 1. There was a slight but significant decrease in relative liver weight in mutant females compared with their wild-type littermates. All other relative organ weights were not significantly different. A gross histopathological examination of various organs was performed. There were no obvious pathological abnormalities in mutant animals (data not shown). Ovaries from mutant and wild-type animals were indistinguishable and contained follicles from all stages of development, as well as corpora lutea (Fig. 3, a and b). Mammary glands from adult virgin wild-type and mutant mice showed normal ductal development and end bud formation (Fig. 3, c and d). Involuting mammary glands from parous wild-type and mutant females did not show any differences (Fig. 3, e and f). There were also no histological differences in uteri from wild-type and mutant animals (Fig. 3, g–j). The photographs in Figure 3, g and h, were taken at a different stage of the estrous cycle compared with the higher magnifications of uterine tissue shown in Figure 3, i and j. These histological findings, as well as the indistinguishable relative uterine weights in wild-type and mutant animals (Table 1), did not support the hypothesis that Gpr30 might be involved in the endocrine regulation of reproductive function.
Histological analysis of reproductive organs from wild-type and mutant mice. Representative organs of adult wild-type (+/+, n = 9) and mutant littermates (−/−, n = 8) are shown. In the ovaries (a and b), follicles from all stages, as well as corpora lutea, were present in both genotypes. Mammary glands from adult virgin mice (c and d), and from adult parous mice (e and f) were indistinguishable. g–j) There were also no differences in uterine morphology between wild-type and mutant animals. Uteri depicted at lower magnification (g and h) were from a different estrous cycle stage than uteri depicted at higher magnification (i and j). Original magnification ×30 (a, b, g, and h), ×20 (c–f), and ×60 (i and j).
To further elucidate the functional involvement of GPR30-mediated signaling in male and female reproduction, we performed continuous mating studies. Reproductive outcome was examined by determining the number of litters and offspring produced by wild-type and mutant mice. In toto, each male was mated with four females in the 2-mo mating period. Wild-type males sired 3.9 ± 0.3 litters and 6.6 ± 0.8 offspring per litter, and mutant males sired 3.8 ± 0.7 litters and 6.3 ± 0.9 offspring per litter (not significantly different from the litter size of wild-type males). Each female was mated twice in the 2-mo mating period. Wild-type females produced an 1.9 ± 0.4 litters and 8.2 ± 1.5 offspring per litter. Mutant females produced 1.7 ± 0.5 litters and 7.5 ± 1.2 offspring per litter (not significantly different from the litter size of wild-type females). These results demonstrated that the absence of GPR30 protein had no major effect on male or female fertility.
To analyze whether Gpr30-deficient females exhibited more subtle defects in uterine responses after estradiol stimulation, we ovariectomized wild-type and mutant females and treated them 14 days after ovariectomy for 3 days with vehicle or estradiol. We analyzed uterine epithelial cell proliferation, increases in relative uterine weight and uterine epithelial cell height, and the induction of estradiol target genes (Figs. 4 and 5). After vehicle treatment, only solitary epithelial cells proliferated in the uteri of wild-type (Fig. 4a) or mutant (Fig. 4c) animals. Estradiol treatment clearly increased uterine epithelial cell proliferation in wild-type (Fig. 4b) and mutant (Fig. 4d) animals. As demonstrated in Figure 5a, quantitative analysis of uterine epithelial cell proliferation in response to vehicle or estradiol treatment did not reveal any significant differences between the two genotypes. Compared with vehicle treatment, estradiol increased relative uterine weights (Fig. 5b) and epithelial cell heights (Fig. 5c) in wild-type and mutant animals to the same extent. Several genes such as Ltf (Fig. 5d), Wnt4 (Fig. 5e), Hspa5 (Fig. 5f), calpactin (S100a10) (data not shown), and cyclin E1 (Ccne 1) (data not shown) were induced after estradiol treatment in uteri from wild-type and mutant animals. The lack of GPR30-mediated signaling had no significant effect on the induction of these genes.
Uterine epithelial cell proliferation in response to estradiol is not impaired in Gpr30-deficient mice. Paraffin sections stained with anti-BrdU antibodies from uteri of vehicle-treated wild-type (a) and mutant (c) animals and from estradiol-treated wild-type (b) and mutant (d) animals are shown. Compared with vehicle treatment, uteri from wild-type and mutant animals showed an increase in epithelial cell height and proliferation after estradiol treatment. Original magnification ×240.
Uterine responses after estradiol stimulation were preserved in Gpr30-deficient mice. Uteri from wild-type animals (white bars) and mutant animals (black bars) exhibited the same characteristic estrogenic responses after estradiol treatment. Epithelial cell proliferation (a), relative uterine weight (b), epithelial cell height (c), and expression of the estrogen-responsive genes Ltf (d), Wnt4 (e), and Hspa5 (f) were analyzed. ***P(treatment) < 0.001.
To investigate whether Gpr30 was involved in the mediation of estrogenic responses in the mammary gland, we treated ovariectomized mice for 3 wk with vehicle, estradiol (to mimic puberty-induced changes in the mammary gland), or estradiol plus progesterone (to mimic pregnancy-induced changes in the mammary gland). After vehicle treatment, only a rudimentary ductal system was visible in the mammary fat pad of 6-wk-old wild-type (Fig. 6a) and mutant (Fig. 6d) animals. Estradiol treatment stimulated ductal elongation and end bud growth in wild-type (Fig. 6b) and mutant (Fig. 6e) mammary glands. Combined estradiol plus progesterone treatment enhanced ductal side branching in mammary glands from wild-type (Fig. 6c) and mutant (Fig. 6f) animals. Quantification of end bud formation (Fig. 7a) or ductal side branching (Fig. 7b) did not reveal any significant differences between the two genotypes. The absence of Gpr30 had no effect on end bud formation or side branching in the mammary gland. Ductal epithelial cell proliferation in response to estradiol or combined estradiol-progesterone treatment was unimpaired in mammary glands of Gpr30-deficient mice (Fig. 7c). The Indo gene is a classic estradiol target gene in the mammary gland that can be repressed by progesterone treatment [20]. Remarkably, the absence of Gpr30 had no significant effect on the induction of the Indo gene by estradiol or its repression by combined estradiol-progesterone treatment (Fig. 7d).
Mammary gland responses after hormonal stimulation were unimpaired in GPR30-deficient mice. Ovariectomized wild-type animals (a–c) or mutant animals (d–f) were treated with vehicle (a and d), estradiol (b and e), or estradiol plus progesterone (c and f). Mammary glands of both genotypes showed ductal elongation and end bud formation after estradiol treatment, as well as enhanced ductal side branching after combined estradiol-progesterone treatment. Original magnification ×20.
Mammary gland responses after hormonal stimulation. Ovariectomized wild-type animals (white bars) or mutant animals (black bars) were treated with vehicle, estradiol, or estradiol plus progesterone. End bud formation (a; labeled Endbuds), ductal side branching (b; labeled Sidebranches), mammary epithelial cell proliferation (c), and expression of the estradiol target gene Indo (d) were analyzed. Mammary glands from mutant and wild-type animals responded equally well to the different hormonal treatment paradigms. *P(treatment) < 0.05. ***P(treatment) < 0.001.
We analyzed our overall results using different readout paradigms. We were unable in the uterus or the mammary gland to demonstrate the involvement of GPR30-mediated signaling in reproduction or typical estrogenic responses.
Discussion
Gpr30 has been claimed to act as a novel ER [6, 7]. The expression of Gpr30 in oxytocin-positive neurons [21] and in granulosa cells of preantral follicles [14], as well as the observation that mammary and uterine epithelial cells express Gpr30 in an estrous cycle-dependent manner [13, 14], has led to the speculation that this presumptive ER might mediate estrogen actions in reproductive organs [13, 14, 21].
To test this hypothesis in vivo, we generated Gpr30-deficient mice. Our main findings were as follows: 1) development of reproductive organs was unimpaired in Gpr30-deficient mice, 2) absence of Gpr30 did not impair male or female fertility, and 3) estrogenic responses in the uterus and the mammary gland were completely maintained in Gpr30-deficient animals. These findings are remarkable and unexpected for a receptor that has been claimed to be involved in estradiol signaling [6, 7]. Ovaries developed normally in the complete absence of GPR30. Follicles from all stages of development were present, as well as corpora lutea. This observation and the indistinguishable relative uterine weights in wild-type and mutant females indicated that there was no impairment of the hypothalamic-pituitary-gonadal axis in female mutant mice. The results from continuous mating studies further support this conclusion. The observation that Gpr30-deficient mice have no reproductive impairments is in line with reports from a distinct Gpr30-deficient mouse model in which hinds were given that the mutant mice were fertile [22]. It was recently claimed that Gpr30 is involved in primordial follicle formation in the hamster ovary [23]. However, there are some doubts regarding this hypothesis. First, no specific saturable estradiol binding to GPR30 protein in hamster ovaries was shown [23]. Second, the effects of micromolar estradiol concentrations on calcium currents in hamster ovarian cells were negligible [23]. Third, the fact that 100 nM ICI was unable to compete with micromolar estradiol (coupled to BSA) in ex vivo studies does not allow the conclusion that Gpr30 is involved in primordial follicle formation. With respect to estradiol, ICI was underdosed. Therefore, we believe that these ex vivo findings with hamster ovaries [23] and our results, as well as hinds from other Gpr30-deficient mouse models [22], lead to the conclusion that evidence for a functional role of GPR30 in ovarian and reproductive physiology in vivo is lacking.
An alternative explanation for our negative results would be that Gpr30 was not completely inactivated in our mouse model. However, we can rule out that possibility. According to our targeting strategy, we eliminated exon 3 of the Gpr30 gene and, thus, the complete open reading frame of the GPR30 protein. As expected from this strategy, the expression of Gpr30 mRNA was completely absent in the mutant animals. Most important, we also removed the selection marker from the murine genome before performing the phenotypic analysis. It was previously demonstrated that neomycin resistance cassettes could silence genes hundreds of kilobytes away from the targeted gene and, thus, produced phenotypes that were independent of the targeted gene [24].
Apart from analyzing reproductive behavior, we also performed experiments that would allow the identification of more subtle deficits in estrogenic responses in the uterus and the mammary gland. As shown by our experiments, Gpr30 was expressed in all reproductive organs, and it was previously demonstrated that in Esr1-deficient mice genes such as Hspa5, S100a10, and Wnt4 were still inducible by estradiol and were not blocked by the classic ER antagonist ICI182780 [3, 4]. Therefore, it was speculated that other G protein-coupled ERs might exist [3, 4]. In our study, we could not demonstrate that Gpr30 was involved in the mediation of classic uterine responses to estradiol stimulation such as increases in epithelial cell height or uterine weight and cellular proliferation. The lack of effects on uterine epithelial cell proliferation in the case of the absence of Gpr30 underlined that Gpr30 was not involved in vivo in the transactivation of the EGFR pathway, resembling a nongenomic mechanism of estradiol action in normal uterine physiology. Moreover, GPR30 did not seem to have a role for the induction of uterine genes such as Hspa5, S100a10, and Wnt4, which were claimed to be induced by estradiol in an Esr1-independent manner [3, 4]. It was recently speculated that ER beta (Esr2) might inhibit Esr1-mediated proliferation of uterine epithelial cells [25]. Knockin mice harboring a point mutation in the ligand-binding domain of Esr1 were completely unresponsive to estradiol, but synthetic ER alpha-selective agonists such as propyl pyrazole triol (PTT) and diethylstilbestrol (DES) were still active on this mutant ER. Exceedingly high doses of DES compared with PTT were required to stimulate uterine growth in these animals. Because DES induced uterine Esr2 levels and was able to activate Esr2, it was speculated that Esr2 might inhibit Esr1-mediated uterine epithelial cell proliferation. One might assume that Gpr30 might also exert such inhibitory effects. However, in our mice we did not observe any evidence for significantly increased estradiol activity in the absence of Gpr30 in any of the readouts examined. Therefore, we rule out the possibility that Gpr30 might have inhibitory activity on Esr1-mediated uterine effects. In line with this, we were unable to demonstrate inhibitory effects of the Gpr30-selective agonist G1 [26] on estradiol-activated responses in the uterus [27]. Because our data were obtained from normal uteri and mammary glands, we cannot rule out the possibility that the absence of Gpr30 might have an effect on the initiation, development, and growth of tumors in mice.
According to our analysis, Gpr30 was dispensable for mammary gland development and had no effect on mammary epithelial cell proliferation. Ductal elongation, end bud formation, and ductal side branching were maintained in Gpr30-deficient animals. Several of these estrogenic responses in the mammary gland are mediated via downstream EGFR signaling [9, 10]. As evidenced by the phenotypes of Esr1-deficient mice [2] and Gpr30-deficient mice, Esr1 (but not Gpr30) is the ER involved in these processes. Notably, Esr1 can transactivate the EGFR pathway [11]. Whereas Esr1-deficient mice were infertile and showed impaired ductal mammary gland development during puberty and had hypoplastic uteri that were refractory toward estradiol stimulation [2], GPR30 did not seem to have a role in these processes. Aromatase (Cyp19a1)-deficient mice display an arrest of follicular development at the antral stage and are infertile because of lack of ovulation [28]. On the other hand, these strong discrepancies between the phenotypes of Gpr30-deficient and Esr1-deficient or Cyp19a1-deficient mice raise doubts regarding the potential role of Gpr30 as an ER. Starting with two publications reporting that GPR30 protein binds to and signals in response to estradiol [6, 7], there are now many subsequent studies that (without performing the critical experiments) claim that Gpr30 is responsible for any estrogenic effect that cannot be blocked by ICI or that can be activated by the GPR30-selective agonist G1 [26]. However, it has to be taken into account that G1 may have effects unrelated to GPR30, as well as that estradiol effects which cannot be blocked by ICI182780 (such as the induction of Hspa5, S100a10, and Wnt4 in the uterus after estradiol application) are not all GPR30 mediated, as demonstrated in our study. Most important, there are many conflicting observations regarding the role of Gpr30 as an ER. Although Filardo and colleagues [7] reported that SKBR3 cells (expressing GPR30 protein) bind to estradiol and signal in response to estradiol, Pedram and coworkers [29] could demonstrate neither specific estradiol binding to SKBR3 cells nor nongenomic estradiol signaling in these cells. Silencing of Gpr30 in Esr1-positive MCF-7 cells had no effect on their nongenomic signaling abilities [29], and the only molecular moiety that bound estradiol in MCF-7 cells was ESR1 (but not GPR30) [29]. We could demonstrate neither saturable specific binding of radioactive estradiol to GPR30 nor cAMP or calcium signaling in GPR30-positive cells in response to estradiol stimulation [27]. In addition, we failed to demonstrate that the GPR-30-selective agonist G1 [26] stimulated any estrogenic effects in the uteri or mammary glands of ovariectomized mice [27]. An earlier study [30] mentioned that estradiol had no effect on the growth inhibitory effects of Gpr30 on MCF-7 cell proliferation. From these data, it was concluded that Gpr30 might be a true orphan receptor or that its ligand was present in the serum that was used to culture the cells or was even secreted by MCF-7 cells [30]. In addition, the phylogenetic relationship of Gpr30 to peptide receptors [5] (such as the interleukin 8 receptor and the angiotensin receptor) raises an important caveat against its potential role as an ER.
Taken together, we conclude that Gpr30 is dispensable for the normal development of reproductive organs and reproductive function in mice. Gpr30 does not mediate estrogenic responses in the uterus and the mammary gland. We anticipate that these findings will stimulate further research, leading to the deorphanization of this G protein-coupled receptor that is still in search of a physiological role.
