Characterization, Neurosteroid Binding and Brain Distribution of Human Membrane Progesterone Receptors δ and ϵ (mPRδ and mPRϵ) and mPRδ Involvement in Neurosteroid Inhibition of Apoptosis

Abstract

Three members of the progestin and adipoQ receptor (PAQR) family, PAQR-7, PAQR-8, and PAQR-5 [membrane progesterone (P4) receptor (PR) (mPR)α, mPRβ, and mPRγ], function as plasma mPRs coupled to G proteins in mammalian cells, but the characteristics of two other members, PAQR6 and PAQR9 (mPRδ and mPRϵ), remain unclear, because they have only been investigated in yeast expression systems. Here, we show that recombinant human mPRδ and mPRϵ expressed in MDA-MB-231 breast cancer cells display specific, saturable, high-affinity [³H]-P4 binding on the plasma membranes of transfected cells with equilibrium dissociation constants (K_ds) of 2.71 and 2.85 nm, respectively, and low affinity for R5020, characteristics typical of mPRs. P4 treatment increased cAMP production as well as [³⁵S]-guanosine 5′-triphosphate (GTP)γS binding to transfected cell membranes, which was immunoprecipitated with a stimulatory G protein antibody, suggesting both mPRδ and mPRϵ activate a stimulatory G protein (Gs), unlike other mPRs, which activate an inhibitory G protein (Gi). All five mPR mRNAs were detected in different regions of the human brain, but mPRδ showed greatest expression in many regions, including the forebrain, hypothalamus, amygdala, corpus callosum, and spinal cord, whereas mPRϵ was abundant in the pituitary gland and hypothalamus. Allopregnanolone and other neurosteroids bound to mPRδ and other mPRs and acted as agonists, activating second messengers and decreased starvation-induced cell death and apoptosis in mPRδ-transfected cells and in hippocampal neuronal cells at low nanomolar concentrations. The results suggest that mPRδ and mPRϵ function as mPRs coupled to G proteins and are potential intermediaries of nonclassical antiapoptotic actions of neurosteroids in the central nervous system.

Extensive evidence has accumulated over the past decade that progestins exert rapid, cell surface-mediated (i.e. nonclassical) actions that are frequently nongenomic through activation of receptors belonging to the progestin and adipoQ receptor (PAQR) family, named membrane progesterone (P4) receptors (PRs) (mPRs) (1–10). PAQRs are novel 7-transmembrane receptors comprising 11 members, which can be divided into three classes based on their structure and ligand binding characteristics (11, 12). Class II PAQRs consists of five members only present in vertebrates, which includes PAQR7 (mPRα), which was first identified and characterized as a membrane progestin receptor in spotted seatrout (2), PAQR8 (mPRβ) and PAQR5 (mPRγ), which were identified and characterized in humans and other vertebrates (1), as well as PAQR6 and PAQR9, which have been shown to respond to progestins in yeast recombinant expression systems and named mPRδ and mPRϵ, respectively (8).

mPRα is the most extensively characterized progestin membrane receptor and is widely expressed in vertebrate tissues (1, 11, 13, 14). Progestin induction of oocyte meiotic maturation in several teleost species is primarily mediated through mPRα (2, 6, 10, 15), and mPRα is the likely intermediary in progestin stimulation of teleost sperm hypermotility (16, 17). Progestin inhibition of apoptosis in teleost granulosa cells and in human breast cancer cells is also induced through mPRα (18, 19). Interestingly, P4 also reverses the mesenchymal transition of human breast cancer cells through mPRα (5). Moreover, mPRα is likely involved in inhibition of GnRH release from rodent GnRH neurons by a mechanism independent of the nuclear PR (20). More limited studies on mPRβ indicate it has similar steroid binding and signaling characteristics to mPRα (10, 21). Results suggest that P4 induction of oocyte maturation in Xenopus is mediated through this receptor (22) and that mPRβ also has a role, together with mPRα, in progestin stimulation of teleost oocyte maturation (23, 24). These two mPR subtypes also have similar functions in human myometrial cells as intermediaries in P4 transactivation of the nuclear PR (21). mPRγ is widely expressed in reproductive and nonreproductive normal and malignant tissues (1, 11, 18, 25, 26) and is present on the apical membrane of ciliated epithelial cells in fallopian tubes, suggesting it may have a role in regulating ciliary activity and gamete transport (25, 27). Recombinant mPRγs produced in both prokaryotic and eukaryotic expression systems have the binding characteristics of progestin receptors (1, 2, 8), but the signal transduction pathways activated through this putative receptor are unknown. The two remaining class II PAQR members, mPRδ and mPRϵ, respond to progestins in yeast recombinant protein expression systems linked to a β-galactosidase reporter, suggesting they may function as progestin receptors (8). However, essential criteria for designation of mPRδ and mPRϵ as progestin membrane receptors have not been met such as the demonstration that the recombinant proteins produced in homologous expression systems display specific progestin binding characteristic of steroid receptors and that they activate signal transduction pathways in response to progestins.

The hypothesis that recombinant mPRδ and mPRϵ stably expressed on mammalian cell membranes have the steroid binding and second messenger signaling characteristics of membrane progestin receptors was tested in the present study. Although the steroid binding functions of class II PAQRs are now widely accepted, their ability to associate with G proteins and activate them is still disputed (8, 28). Therefore, activation of G proteins by recombinant mPRδ and mPRϵ was also examined to further evaluate this proposed mechanism of signal transduction by mPRs and determine whether is ubiquitous in this class of PAQRs. A distinguishing feature of mPRδ is that its mRNA is exclusively expressed in the human brain, whereas all the other mPRs show a broad tissue distribution (11). Consequently, the ability of mPRδ to bind neurosteroids and act as an intermediary in neurosteroid activation of signaling pathways and modulation of cell functions was also investigated. Finally, the distribution and relative abundances of mPRδ, mPRϵ, and the other mPRs in different regions of the human brain were determined.

Materials and Methods

Chemicals

Chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise noted. Steroids and neurosteroids were purchased from Steraloids (Newport, RI), R5020 was purchased from GE Healthcare (Piscataway, NJ) and Org OD-02-0 and Org OD-13-0 were obtained from Organon (Oss, The Netherlands). [2,4,6,7-³H]-P4 ([³H]-P4, 102 Ci/mmol) and [³⁵S]-GTP (guanosine 5′-triphosphate) γS (1250 Ci/mmol, 12.5mCi/ml) were purchased from GE Healthcare.

Expression of recombinant mPRδ and mPRϵ proteins in MDA-MB-231 cells

Full-length cDNAs encoding human mPRδ (NM_024897) and mPRϵ (NM_198504) were cloned into pCMV6-Neo mammalian expression vectors (OriGene, Rockville, MD) and stably transfected in human breast cancer MDA-MB-231 cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) and selected with 500 μg/ml geneticin (G-418; Invitrogen) after the limited dilution procedures described previously for stable and high expression of mPRα (3, 7). Stable, cell surface expression of mPRs has been achieved previously with this PR-negative cell line (3, 24). Transfected cells were cultured in an atmosphere of 5% CO₂ in DMEM (Sigma-Aldrich) containing 10% fetal calf serum (Invitrogen).

Reverse transcription-polymerase chain reaction

Correct expression of mPR RNAs in transfected cells and in rat hippocampal neuronal cells (H19-7; American Type Culture Collection, Manassas, VA) was confirmed by RT-PCR using gene specific primers for human mPRδ (sense, 5′-CAGATGACCAACGAGACGG-3′ and antisense, 5′-GGAAGTAGCAGATGTGGCG-3′), mPRϵ (sense, 5′-TGCTACAAAGGGATCCCAAC-3′ and antisense, 5′-TGGCACAGATGATTGGAAAA-3′); rat mPRδ (sense, 5′-GTTTTGGGAAGAAGGCATCA-3′ and antisense, 5′-GGCAGTAGGAAGACCAGCAG-3′), and mPRϵ (sense, 5′-CCGGGTCTCTTTGACATCAT-3′ and antisense, 5′-TACTGCAGAATTCGGTGCTG-3′). Total RNA extracted from cells was reverse transcribed into cDNA using SuperScript III, which was used as a template for PCR with the cycling profile of 94 C for 30 sec, 55 C for 30 sec, and 72 C for 1 min for 30 cycles, followed by a 72 C, 10-min extension step.

Plasma membrane preparation

Plasma membranes were prepared as described previously (3) with minor modifications. Cells were homogenized in ice-cold homogenization buffer [25 mm HEPES, 10 mm NaCl, 1 mm dithioerythritol, and 1 mm EDTA (pH 7.6), HAED] containing 1% protease inhibitor cocktail (Pierce, Rockford, IL), and the homogenate was centrifuged for 7 min at 1000 × g, followed by centrifugation of the supernatant at 20,000 × g for 20 min to pellet plasma membranes, which were resuspended in HAED buffer for receptor binding assay or stored at −80 C for future analysis.

Western blot analysis

Western blotting of plasma membranes was conducted as described previously (3) using rabbit polyclonal antibodies generated against human mPRδ- and mPRϵ-specific peptide sequences in their N-terminal regions: mPRδ, CQGGPLEGGTQAKQQ and mPRϵ, RNSHSAASRDPPASC conjugated to hemocyanin (dilution: 1:2000). Actin was used as a loading control and detected using a specific antibody (Abcam, Inc., Cambridge, MA). Antibodies were incubated with their peptide antigens (5 μg peptide/μl antibody) at 4 C overnight to test the specificity of the immunoreactions. In addition, the specificity of each antibody for a single mPR was evaluated by comparing their immunoreactions to plasma membranes of MDA-MB-231 cells overexpressing mPRδ or mPRϵ with those of MDA-MB-231 cells overexpressing the other mPRs.

mPR binding assays

Specific PR binding to plasma membranes of mPRδ- and mPRϵ-transfected MDA-MB-231 cells was assayed following procedures described previously for recombinant mPRα (3). Plasma membranes were incubated for 30 min at 4 C with 4 nm [³H]-P4 alone (total binding), and [³H]-P4 with 100-fold excess nonradioactive P4 (nonspecific binding) in single point assays, bound [³H]-P4 was separated from free by rapid filtration over Whatman GF/B filters using a 36-well cell harvester (Brandel, Gaithersburg, MD) and radioactivity bound to the filters counted in a liquid scintillation counter. Saturation of P4 binding was examined over a range of [³H]-P4 concentrations (0.25-8 nm) and analyzed by nonlinear regression (GraphPad Prism Software, San Diego, CA). Competitive binding assays were conducted in triplicate with steroid competitors (range of 10⁻⁵ to 10⁻¹⁰ m) incubated with 2 nm [³H]-P4 and the results expressed as a percentage of maximum specific binding of P4.

Immunocytochemistry

Localization of recombinant mPRδ and mPRϵ proteins in transfected cells was examined by immunocytochemistry using protocols described previously for mPRα and mPRβ with minor modifications (3, 20). Cells were incubated with specific mPRδ or mPRϵ antibodies (1:500) in PBS containing 2% BSA overnight at 4 C. Cells were then incubated with Alexa Fluor 488 goat antirabbit secondary antibodies (dilution 1:1000; Molecular Probes, Carlsbad, CA) and stained with 4′,6-diamidino-2-phenylindole (0.01% in PBS) to visualize nuclei.

[³⁵S]-GTPγS binding assay

Activation of G proteins by progestin treatment in mPRδ- and mPRϵ-transfected cells was assessed by measuring the increase in specific [³⁵S]-GTPγS binding to plasma membranes (100 nm) after a 20-min incubation as described previously (3).

Immunoprecipitation of G proteins and coimmunopreciptation with mPRs

[³⁵S]-GTPγS-labeled G protein α-subunits activated by progestins in mPR-transfected cells were immunoprecipitated with specific G protein-subunit antibodies [inhibitory G protein (G_i) α and stimulatory G protein (G_s) α, 1:300; Santa Cruz Biotechnology, Inc., Santa Cruz, CA] using protein A/G plus-agarose beads (Santa Cruz Biotechnology, Inc.), and radioactivity was counted as described previously (3). Procedures published previously for coimmunoprecipitation mPRα with G proteins using goat G_s and G_i antibodies (1:100; Santa Cruz Biotechnology, Inc.) and agarose beads were followed (3), and Western blottings were conducted with mPRδ and mPRϵ antibodies as described above.

cAMP assay

cAMP levels were assayed in mPR-transfected cells as a measure of adenylyl cyclase (AC) activity in response to steroid treatments. Equal numbers of cells were seeded in each well, and cells were incubated in serum-free medium overnight to reduce background AC activity before steroid treatment for 15 min and measurement of cAMP in cell lysates using an EIA kit (Cayman Chemical, Ann Arbor, MI) following the manufacturer's instructions.

Quantitative PCR (QPCR)

mRNA levels of mPRs and PR were assayed in human brain tissue cDNA arrays, prepared from pools (2–5) of selected regions of male and female brains of Caucasian adults (age range, 16-65 yr), which included reproductive-aged subjects that had experienced sudden death (HBRT301; OriGene). The manufacturer-prepared Poly (A)+ RNA by oligo (dT) selection to facilitate detection of weakly expressed mRNAs, confirmed its RNA integrity by Northern blot hybridization, and synthesized first strand cDNA using oligo (dT) primers and Maloney murine leukemia virus (MLV) reverse transcriptase. The efficiency of RT was evaluated by confirming the presence of complete transcripts for selected rare and long mRNAs in the cDNA pools. cDNA pools were then normalized so that each contained an equivalent amount of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA. Expression levels of mPRδ, mPRϵ, mPRα, mPRβ, mPRγ, and PR mRNAs were measured by QPCR, using an Eppendorf RealPlex Mastercycler (Eppendorf, Hamburg, Germany) in a 25-μl one-step Brilliant II SYBR Green QPCR Master Mix (Stratagene, La Jolla, CA) containing 100 nm sense and antisense primers. The same mPRδ and mPRϵ primers as those listed above for RT-PCR were used for QPCR. Other human mPR primers were: mPRα sense, 5′-CTGAAGTTTGCCTGACACCA-3′ and antisense, 5′-AATAGAAGCGCCAGGTCTGA-3′; mPRβ sense, 5′-TACCTCACCTGCAGCCTTCT-3′ and antisense, 5′-GCAACAGCCAGCACAAGATA-3′; mPRγ sense, 5′-ACTATGGTGCCGTCAACCTC-3′ and antisense, 5′-TGTACGGATAAGCAAAGGC-3′; and PR sense, 5′-GAGCTTAATGGTGTTTGGTC-3′ and antisense, 5′-GTTTGACTTCGTAGCCCTT-3′. Confirmation that each well of the array plate had equal amounts of cDNA was obtained using a housekeeping gene, GAPDH (sense, 5′-GCCATCAAGGAGGCTGTAAAAGC-3′ and antisense, 5′-GGTATCACCGAGGAAGTCCGTA-3′; OriGene), as a loading control (Supplemental Fig. 1A, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org), and also by examining the manufacturer's validation results. QPCR was performed with a denaturation step of 94 C for 10 min, followed by 40 cycles of 94 C for 30 sec, 55 C for 30 sec, and 72 C for 1 min. The melting curve was plotted after the amplification cycles with the profile of 95 C for 15 sec, 55 C for 15 sec, then ramping the temperature up to 95 C in 20 min and maintaining it for 15 sec to verify the specificity of the PCR products. PCR efficiency of each pair of QPCR primers was tested using serially diluted, reverse transcribed RNA samples obtained from the corresponding mPR-transfected cell line. cDNA levels were calculated from the equation of efficiency (EFF) = 10(−1/slope) − 1. All the primers had similar efficiencies, approximately 97-99%. QPCR data were analyzed using the comparative quantitation method, the relative expression level of each gene was calculated from 2^−ΔCt (2^{−(Ct mPR-Ct GAPDH)}) values and results multiplied by 10⁴ and defined as arbitrary units in the figure.

MAPK and serine-threonine kinase (AKT) activation assay

Increases in ERK and AKT phosphorylation in response to progestin treatments were measured in mPRδ-transfected MDA-MB-231 cells that had been serum starved for 4-5 d by Western blot analysis using monoclonal antibodies directed against total p42/44 (ERK) and phospho-p42/44 (p-ERK) (32), or polyclonal antibodies directed AKT and p-AKT (Cell Signaling, Danvers, MA).

Cell death and terminal deoxynucleotidyl transferase 2′-deoxyuridine, 5′-triphosphate nick end labeling (TUNEL) assays

Cells (70% confluent) were cultured with steroids in serum-free DMEM for 4 d, harvested, stained with 15% trypan blue solution, and approximately 500 cells from each treatment group were examined for trypan blue exclusion under a microscope to calculate percent cell death. A TUNEL assay kit (Promega, Madison, WI) was used to label fragmented DNA of mPRδ-transfected cells and H19-7 rat hippocampal neuronal cells, and the proportion of apoptotic nuclei was counted under a fluorescent microscope in 10 random fields totaling approximately 500 cells.

Statistics

Data presented are means ± sem from at least three observations, and all experiments were repeated three or more times on different days. Significant differences between paired treatment groups were analyzed by paired Student's t test and multiple treatment groups by one-way ANOVA and Tukey's (GraphPad Prism Software).

Results

Expression and localization of recombinant mPRδ and mPRϵ in MDA-MB-231 cells

Bands of the expected size showing increased expression of mPRδ and mPRϵ mRNAs in transfected cells were detected by RT-PCR, whereas no mPRδ transcripts and lower expression of mPRϵ were detected in control vector-transfected cells (Figs. 1A and 2A, left). Western blot analysis with mPRδ and mPRϵ antibodies showed strongly stained approximately 80-kDa bands, likely representing mPR dimers that are commonly observed with other mPRs (3, 20, 21, 25, 29), on plasma membranes of mPRδ- and mPRϵ-transfected cells, but not in control vector-transfected cells (Figs. 1A and 2A, right). Immunocytochemical analysis of transfected cells showed positive fluorescent staining of both recombinant proteins, which were primarily localized on plasma membranes (Figs. 1B and 2B), whereas no fluorescent signals were detected in vector-transfected cells. The specificities of the antibodies were confirmed by blocking the immunoreactions with the peptide antigens (Figs. 1, A and B, and 2, A and B). The mPRδ and mPRϵ antibodies are specific for their respective mPR subtype, because no immunoreactive bands were detected using either of them in cells overexpressing other mPRs (Supplemental Fig. 1, C and D). The results indicate successful expression of human mPRδ and mPRϵ mRNAs and proteins in stably transfected MDA-MB-231 cells.

Fig. 1.

Expression, localization, and ligand binding characteristics of recombinant mPRδ in stably transfected MDA-MB-231 cells. A and B, Expression of mRNA (left), recombinant protein on cell membrane (right) (A) and immunocytochemical analysis (B) of mPRδ-transfected (mPRδ) cells. Mkr, Molecular weight marker; Vec, empty vector-transfected cells; P, peptide antigen. C, Single point-specific [³H]-P4 binding to plasma membranes from Vec and mPRδ cells. D, Representative saturation analysis and Scatchard plot of [³H]-P4 binding to mPRδ cell membranes. E and F, Representative competition curves of steroid binding to mPRδ cell membranes expressed as a percentage of maximum [³H]-P4 binding. T, Testosterone; E2, 17β-estradiol; Cort, cortisol; 02, Org OD-02-0; 13, Org OD-13-0; Mkr, DNA size markers. *, P < 0.001 compared with Vec, n = 6.

Fig. 2.

Expression, localization, and ligand binding characteristics of recombinant mPRϵ in stably transfected MDA-MB-231 cells. A and B, Expression of mRNA (left), recombinant protein on cell membrane (right) (A) and immunocytochemical analysis (B) of mPRϵ-transfected cells (mPRϵ). C, Single point-specific [³H]-P4 binding to plasma membranes from control (Vec) and mPRϵ cells. D, Representative saturation analysis and Scatchard plot of [³H]-P4 binding to mPRϵ cells. E and F, Representative competition curves of steroid binding to mPRϵ cell membranes expressed as a percentage of maximum [³H]-P4 binding. Mkr, DNA size markers; P, peptide antigen; T, testosterone; E2, 17β-estradiol; Cort, cortisol; 02, Org OD-02-0; 13, Org OD-13-0; 5020, R5020. *, P < 0.001 compared with Vec, n = 6.

[³H]-P4 binding characteristics of mPRδ and mPRϵ

Significant increases in specific [³H]-P4 binding to plasma membranes of mPRδ- and mPRϵ-transfected MDA-MB-231 cells compared with vector-transfected cells were detected in single point receptor binding assays (Figs. 1C and 2C). Saturation analysis and Scatchard plotting demonstrated high-affinity, limited-capacity, single, specific [³H]-P4 binding sites on plasma membranes of mPRδ- and mPRϵ-transfected cells, with equilibrium dissociation constants (K_ds) of 2.71 ± 0.26 and 2.85 ± 0.33 nm and maximum binding capacities (B_maxs) of 0.70 ± 0.043 and 0.63 ± 0.067 nm/mg protein, respectively (Figs. 1D and 2D). Although a low amount of specific binding was also detected in vector-transfected cells, saturation analysis demonstrated that [³H]-P4 binding was not saturated with concentrations of tracer up to 10 nm, and a linear plot could not be obtained by Scatchard analysis so the binding constants cannot be calculated, indicating that the binding is not due to presence of a receptor protein (Supplemental Fig. 1, E and F). Competitive binding assays showed that P4 had the highest binding affinity to both receptors, testosterone displayed approximately 1/10th the affinity of P4, whereas cortisol and 17β-estradiol showed negligible or no binding to mPRδ and mPRϵ (Figs. 1E and 2E). The mPRα-specific progestin agonists, Org OD-02-0 and Org OD-13-0, showed high binding affinities to both receptors, whereas the nuclear PR agonist R5020 displayed low binding affinity, approximately 1/1000 that of P4, to both receptors (Figs. 1F and 2F).

G_s activation, cAMP signaling, and G_s coupling to mPRδ and mPRϵ

Treatment with P4 (100 nm) significantly increased specific [³⁵S]-GTPγS binding to plasma membranes of mPRδ- and mPRϵ-transfected cells but not to control vector-transfected cells (Fig. 3, A and D), indicating the receptors are intermediaries in P4-induced G protein activation. Immunoprecipitation of membrane-bound [³⁵S]-GTPγS on mPRδ- and mPRϵ-transfected cells with specific G protein α-subunit antibodies showed that most of the elevated GPTγS after P4 treatment was bound to a G_s α-subunit and practically none with the G_i α-subunit (Fig. 3, B and E, bottom). Cellular cAMP levels in both mPRδ- and mPRϵ-transfected cells were significantly elevated after treatment with P4 and the specific mPR agonists, Org OD-02-0 and Org OD-13-0 (Fig. 3, C and F), indicating that AC activity is increased after P4 treatment through these mPRs, which is consistent with an involvement of both mPRs in activation of stimulatory G proteins.

Fig. 3.

G protein activation and signal transduction of recombinant mPRδ and mPRϵ. A and D, [³⁵S]-GTPγS binding to plasma membranes of mPRδ-transfected (A) and mPRϵ-transfected (D) (mPRδ and mPRϵ) cells. Vec, Control vector-transfected cells; Veh, vehicle control. *, P < 0.05 compared with Veh, n = 6. B and E, Immunoprecipitation of activated G protein on plasma membranes of mPRδ (B) and mPRϵ (E) cells. IgG, Control IgG; Gs, anti-Gα_s-subunit antibody; Gi, anti-Gα_i-subunit antibody. *, P < 0.05 compared with IgG, n = 6. Western blot analyses show immunoprecipitated G proteins probed with mPRδ (B) and mPRϵ (E) antibodies. C and F, Effects of 15-min progestin treatments on cAMP concentrations in mPRδ (C) and mPRϵ (F) cells. *, P < 0.05 compared with Veh. n = 6.

Coimmunoprecipitation studies showing bands of the predicted size for mPRs were detected with the mPRδ and mPRϵ antibodies in membrane fractions containing receptor/G protein complexes precipitated with the G_s antibody from mPRδ- and mPRϵ-transfected cells (Fig. 3, B and E, top), providing further evidence that both these mPRs are coupled to stimulatory G proteins.

Distribution and relative expression of mPRs and PR mRNAs in the human brain

QPCR of human brain tissue arrays showed distinctly different expression patterns of mPRδ, mPRϵ, mPRα, mPRβ, mPRγ, and PR mRNAs in the various brain regions (Fig. 4, A and B). The relative expression of mPRδ mRNA was the highest among all the mPRs and PR in most regions of the brain, with highest expression in the corpus callosum, hypothalamus, and spinal cord (Fig. 4, A and B). mPRδ expression was also higher than that of other mPRs in all neocortex lobes, the limbic system (amygdala, hippocampus, and nucleus accumens) and thalamus, and in brain regions important for memory and movement (caudate, putamen), reward (substantia nigra), and autonomic functions (medulla and pons). Highest expression of mPRϵ mRNA was observed in the pituitary gland and hypothalamus (Fig. 4A), and relatively high levels were found in several brain regions where mPRδ showed high expression, such as the limbic system, caudate nucleus accumens, pons, and olfactory bulb (Fig. 4B). Expression of mPRβ mRNA was also high in many of the same brain regions as mPRδ, including the corpus callosum, hippocampus, hypothalamus, substantia nigra, cerebellum, and spinal cord (Fig. 4B). Overall, the relative expression of mPRβ was the second highest in many brain regions and was higher than that of other mPRs in parts of the cerebellum. The expression of mPRα mRNA was lower in many regions of the brain than that of mPRδ, mPRϵ, and mPRβ. For example, mPRα mRNA expression was lower than that of mPRβ in the cerebral cortex and hypothalamus, similar to recent findings in rat brains, where mPRβ is highly expressed in select neurons but not in glial cells (30). Highest mPRα levels were detected in the temporal lobe, pituitary gland, medulla, and spinal cord (Fig. 4B). This mPR subtype is expressed in most neurons, astrocytes and oligodendrocytes in the spinal cord (31). mPRγ showed lower expression than that of other mPRs in many brain regions, and highest relative levels were observed in spinal cord, choroid plexus, and pons (Fig. 4B). Expression of the nuclear PR mRNA was lower than that of all the mPRs in most brain regions but was higher than that of other receptors in the pituitary gland. These results suggest that progestins exert their actions throughout the human brain through mPRs and that different mPRs are the primary mediators of progestin actions in different brain regions.

Fig. 4.

Relative mRNA expression levels of mPR subtypes and PR in different regions of human brain determined by QPCR. A, relative expression of mPRδ and mPRϵ mRNAs in selected brain regions. B, Expression levels of mPRδ, mPRϵ, mPRα, mPRβ, mPRγ, and nPR mRNAs in different human brain regions shown with same arbitrary units as in graph A, Expression scale (arbitrary units): −, ≤1; +, >1 to ≤5; ++, >5 to ≤10; +++: >10 to ≤30; ++++, >30 to ≤60; +++++, >60 to ≤150; ++++++, >150. F. L., Frontal lobe; T. L., temporal lobe; O. L., occipital lobe; P. L., parietal lobe; Pa. G., paracentral gyrus; Po. G., postcentral gyrus; O. B., olfactory bulb; Thal, thalamus; C. C., corpus callosum; Hypo, hypothalamus; Amyg, amygdala; Hippo, hippocampus; Caud, caudate; Puta, putamen; S. N., substantia nigra; P. G., pituitary gland; C. G., cerebellum gray; C. W., cerebellum white; C. V., cerebellum vermis; N. A., nucleus accumbens; Pons, pons; Medu., medulla; S. C., spinal cord; C. P., choroid plexus.

Neurosteroid binding to mPRδ, mPRϵ, mPRα, and mPRβ

Competitive binding studies with four neurosteroids, dehydroepiandrosterone (DHEA), pregnanolone, pregnenolone, and allopregnanolone, showed that all of them competed with [³H]-P4 binding to mPRδ, mPRϵ, mPRα, and mPRβ, with relative binding affinities (RBAs) ranging from approximately less than 1 to 34% that of P4 (Fig. 5, A–D, and Table 1). Pregnenolone displayed the weakest binding affinity for the receptors with IC₅₀s in the micromolar range, followed by pregnanolone and DHEA, which had low affinities for all the receptors except mPRδ. In contrast, allopregnanolone showed relatively high binding affinities for mPRα, mPRβ, and mPRδ with IC₅₀s in the nanomolar range and RBAs 5-34% that of P4 (Table 1). Neurosteroid binding specificity profiles of the four mPRs showed marked differences. mPRϵ displayed greatest binding with pregnenolone but weakest binding with allopregnanolone (Fig. 5D and Table 1). In contrast, mPRβ showed lowest affinity for DHEA and pregnanolone, intermediate affinity for allopregnanalone, and high affinity for P4 (Fig. 5B and Table 1). Interestingly, the brain-specific mPR subtype, mPRδ, showed the highest binding affinities for allopregnanolone, pregnanolone, and DHEA, with IC₅₀ values 37, 11, and 17%, respectively, of the lowest IC₅₀s reported for any of the other mPRs (Fig. 5C and Table 1). Allopregnanolone was the most effective competitor for P4 binding to mPRδ-transfected cell membranes, with an IC₅₀ of 151 nm, one third that of P4. These results suggest that mPRδ may be an important intermediary in neurosteroid actions in the brain.

Fig. 5.

Representative competition curves of neurosteroid binding to cell membranes of mPRα-transfected (A), mPRβ-transfected (B), mPRδ-transfected (C), and mPRϵ-transfected (D) cells expressed as a percentage of maximum [³H]-P4 binding. DHEA, Dehydroepiandrosterone; Pregna, pregnanolone; Preg, pregnenolone; Allo, allopregnanolone.

Table 1.

Comparison of binding affinities of neurosteroids to mPRs

	P4		Dehydroepiandrosterone		Pregnanolone		Pregenolone		Allopregnanolone
	IC50	RBA	IC50	RBA	IC50	RBA	IC50	RBA	IC50	RBA
mPRα	26.3 ± 5.9	100	4541 ± 427	0.58	3139 ± 415	0.84	n/m	n/m	404 ± 73	6.51
mPRβ	29.2 ± 4.5	100	47930 ± 6285	0.06	11180 ± 1190	0.26	n/m	n/m	559 ± 89	5.22
mPRδ	50.8 ± 17.3	100	781 ± 71	6.50	346 ± 69	14.68	2943 ± 336	1.73	151 ± 47	33.64
mPRϵ	101.6 ± 20.1	100	n/m	n/m	3296 ± 326	3.10	1272 ± 201	8.00	2561 ± 505	4.00

Each value is the mean of at least three separate competitive binding assays. IC₅₀ is the competitor concentration (nm) that causes 50% displacement of [³H]-P4. RBA is the binding affinity compared with that of P4 in percent. n/m, IC₅₀ not measurable.

Effects of allopregnanolone on cAMP levels, ERK, and Akt phosphorylation in mPRδ-transfected cells

Allopregnanolone (20 and 100 nm) caused a dose-dependent significant increase in cAMP levels in mPRδ-transfected cells, mimicking the effects of P4 at these concentrations (Fig. 6A). Phosphorylation of ERK increased after 15 min of P4 treatment in mPRδ-transfected cells (Supplemental Fig. 1B). Allopregnanolone (20 and 100 nm) also increased phosphorylation of ERK levels in mPRδ-transfected cells after 20 min of treatment, but both steroids were ineffective in phosphorylating ERK in vector-transfected cells (Fig. 6B). Allopregnanalone and P4 treatments did not cause phosphorylation of Akt in either mPRδ- or vector-transfected cells (Fig. 6B).

Fig. 6.

Effects of P4 and allopregnanolone (Allo) on second messengers and cell death in mPRδ-transfected (mPRδ), vector (Vec), and hippocampal neuronal cells. A, Effects of 15-min treatments with P4 or Allo and forskolin (Fors) on cAMP concentrations in mPRδ cells. *, P < 0.05; **, P < 0.01 compared with respective vehicle (Veh) control. n = 6. B, Representative Western blot analyses of effects of 20-min treatments with P4 or Allo on activation of ERK 1/2 (left) and Akt (right) in mPRδ and Vec cells. EGF, Epidermal growth factor control; P-ERK, phosphorylated ERK; P-Akt, phosphorylated Akt. C, Effects of P4 and Allo on cell death of Vec and mPRδ cells. *, P < 0.01 compared with Veh control, n = 6. D, Antiapoptotic effects of P4 and Allo in mPRδ or Vec cells (detected with TUNEL assay). E, RT-PCR of mPRδ and mPRϵ mRNAs and antiapoptotoic effects of P4, Allo, and R5020 in rat hippocampal neuronal (H19-7) cells. **, P < 0.001, compared with Veh control, n = 6.

Effects of allopregnanolone and P4 on serum starvation-induced cell death and apoptosis in mPRδ-transfected cells and in hippocampal neuronal cells

Progestins decrease cell death and inhibit apoptosis through mPRα in several vertebrate cell types (18, 19). To determine whether mPRδ has a similar function and has a potential role in the neuroprotective effects of neurosteroids, the effects of P4 and allopregnanolone on serum starvation-induced cell death were investigated in mPRδ-transfected cells. Both allopregnanolone (100 and 500 nm) and P4 (20 and 100 nm) significantly inhibited serum starvation-induced cell death in mPRδ-transfected cells but not in control vector-transfected cells (Fig. 6C). Both steroids were also effective in inhibiting apoptosis at concentrations of 20 nm in the mPRδ-expressing cells but not in vector-transfected controls (Fig. 6D). Similar antiapoptotic effects of P4 and allopregnanolone were observed in a rat hippocampal neuronal cell line (H19-7) that has high expression of mPRδ and mPRϵ mRNAs (Fig. 6E), as well as mPRα and mPRβ, but relatively low expression of PR (results not shown).

Discussion

Three members of class II PAQRs, PAQR7 (mPRα), PAQR8 (mPRβ), and PAQR5 (mPRγ), function as membrane progestin receptors in a variety of vertebrate models and in eukaryotic and prokaryotic expression systems (1–4, 6–8, 10, 19, 22, 24, 33). The present results clearly demonstrate that the two remaining class II PAQRs, PAQR6 (mPRδ) and PAQR9 (mPRϵ), produced in a mammalian recombinant expression system, also display all the P4 binding, G protein activation and signaling characteristics of membrane progestin receptors. The PR binding functions of these recombinant proteins produced in a homologous expression system are consistent with previous observations of Smith et al. (8) that both mPRδ and mPRϵ produced in a yeast recombinant protein expression system linked to a β-galactosidase reporter respond to progestins. However, the ability of mPRs to activate G proteins was questioned by these researchers based on the observation that recombinant mPRs are able to respond to progestins in the absence of G proteins in the yeast system (8). The significance of their findings in a heterologous expression system and their relevance to the normal physiology of mPRs, which are restricted to the vertebrates, is uncertain. On the other hand the present results showing that two additional mPRs, mPRδ and mPRϵ, also activate G proteins in mammalian cells and are coimmunoprecipitated with G proteins provides further evidence that the mPRs are coupled to G proteins and signal through them. G proteins are associated with mPRα and mPRβ in a wide variety of vertebrate cells and are activated through these mPRs in response to progestins (3, 7, 16–21, 29, 33). Thus, all the available evidence obtained to date indicates that the normal signal transduction pathway of the five mPRs (class II PAQRs) in vertebrate cells is through G proteins.

The progestin receptor characteristics of recombinant mPRδ and mPRϵ are very similar to those of the other mPRs. Both of them are expressed on plasma membranes of transfected cells and display high binding affinities for P4, with dissociation constants of approximately 3 nm, in the same range as those of other mPRs (3). Their steroid specificities are also similar to those of the other mPRs, with high binding affinities for P4, moderate binding for testosterone, and low or no binding affinity for other classes of steroids. Interestingly, both receptors showed low RBAs (<1%) to the nuclear progestin agonist, R5020, and high RBAs to the mPRα agonists, Org OD-02-0 and Org OD-13-0, which suggests that the ligand structure requirements for binding to mPRδ and mPRϵ are similar to those of the well-characterized mPRα (3, 32). However, in marked contrast to nearly all the studies with mPRα and mPRβ indicating they activate G_is (3, 7, 18–21, 29), immunoprecipitation of P4-induced [³⁵ S]-GTPγS binding to recombinant mPRδ and mPRϵ cell membranes with specific Gα-subunit antibodies showed that these receptors instead activate stimulatory G proteins. This conclusion is supported by experiments showing that both receptors are coimmunoprecipitated with the G_s antibody and that P4 increases cAMP production in mPRδ- and mPRϵ-transfected cells.

A major finding of this study is that allopregnanolone is an effective competitor for [³H]-P4 binding to mPRδ, displacing 50% [³H]-P4 binding at a concentration of 151 nm. The RBA of allopregnanalone for mPRδ is 33.6% that of P4, whereas it is much lower for mPRα and mPRβ (5-6%). The results show that allopregnanolone acts as a potent mPR agonist in cells transfected with mPRδ, mimicking the stimulatory effects of P4 on cAMP production at low concentrations of 20 and 100 nm. This brain-specific mPR subtype appears to be the primary mPR mediating neurosteroid actions, because it also has a much greater binding affinity than the other mPRs for DHEA and pregnanolone, with IC₅₀s in the nanomolar range. Allopregnanolone and these other neurosteroids could potentially exert their actions via this mPR subtype throughout the brain, because mPRδ shows high expression in all brain regions except the choroid plexus. Moreover, brain regions where allopregnanolone biosynthesis is greatest, the neocortex and hippocampus (34), coincides with those where mPRδ mRNA expression is highest. Allopregnanolone has been shown to act in the brain by binding to γ-aminobutyric acid (GABA)_A receptors and through noncompetitive interactions with N-methyl-d-aspartic acid receptors, resulting in their allosteric modulation (35, 36). Allopregnanolone does not bind to nuclear PRs and is commonly thought to be devoid of any hormonal activity through binding to steroid receptors (37–39). The present results challenge this prevailing opinion and provide the first demonstration that allopregnanolone can bind to and activate steroid receptors. They also indicate an additional pathway, by which the neurosteroid could potentially influence brain functions, through activation of mPRδ and other mPRs, such as mPRβ.

Neuroprotection is considered one of the most important functions of allopregnanolone and other neurosteroids in the brain (34, 36). Allopregnanolone levels decline in patients with neurodegenerative diseases, such as Alzheimer's disease and multiple sclerosis (40). Allopregnanolone exerts neuroprotective effects by a variety of mechanisms, including inhibition of neuronal apoptosis by a nongenomic mechanism involving positive allosteric modulation of GABA_A receptors and negative allosteric modulation of N-methyl-d-aspartic acid receptors. The present results suggest that neuroprotective actions of allopregnanolone and P4 can also be mediated through mPRs. Progestins exert antiapoptotic actions on granulosa and breast cancer cells, inhibiting serum starvation-induced cell death through a mPRα-dependent mechanism (18, 19). The finding that MAPK and Akt are activated in both cell types after treatment with the mPR-specific agonist, Org OD-02-0, is consistent with these antiapoptotic functions of mPRα, because both MAPK and Akt inhibit the proapoptotic gene, Bcl-2-associated death promoter (BAD), and MAPK up-regulates antiapoptotic members of the apoptosis regulator protein Bcl-2 family (41, 42). The current results show that both P4 and allopregnanolone activate MAPK and inhibit cell death and apoptosis in mPRδ-transfected cells. Similar apoptotic actions of allopregnanolone at a low physiological concentration of 20 nm were also observed in hippocampal neuronal cells. This concentration is physiologically relevant, because allopregnanolone levels are nearly twice as high (35 nm) in mouse brains during the dark phase (scotophase) of the diurnal cycle and two to five times higher in the plasma of women during the third trimester of pregnancy (43–45). In contrast, modulation of GABA_A receptor functions are typically observed with micromolar concentrations of allopregnanolone (46, 47). Previous studies showing that some neuroprotective effects of P4 in rat hippocampal neurons and of allopregnanolone in a rat PR-negative, mPR-positive neuronal cell line (GT1-7 cells) are mediated through MAPK-dependent pathways (9, 48) are consistent with our results. However, the finding that progestins do not activate Akt in mPRδ-transfected cells differs from previous observations with mPRα in granulosa and breast cancer cells (18, 19) and indicates that different second messenger pathways mediate antiapoptotic actions of progestins through mPRδ and mPRα, which may be a consequence of them being coupled to different G proteins. In conclusion, the present results are consistent with a proposed neuroprotective role for mPRs, but direct involvement of mPRs in mediating some of the neuroprotective effects of allopregnanolone in neuronal cells remains to be investigated.

The significance of the multiplicity of class II PAQR subfamily remains a major unresolved issue in understanding their physiological roles. Although multiple subtypes have been identified for other steroid receptors, the finding that five distinct receptor subtypes mediate the actions of a single class of steroids is unprecedented. The observation that multiple mPRs are coexpressed in all the mammalian cell types investigated to date (7, 9, 18, 20, 21, 27, 29) suggests the presence of a multifaceted nonclassical progestin signaling mechanism in these cells. This complexity of progestin signaling through multiple mPRs complicates the investigation of their individual physiological roles. The present results provide the first indication of the possible functional significance of multiple mPR subtypes. On the basis of the binding studies, it is predicted that mPRα and/or mPRβ would be preferentially activated by P4 resulting in activation of an G_i, whereas mPRδ would be preferentially activated by allopregnanolone, DHEA and pregnanolone causing activation of G_s. Activation of mPRϵ would also occur at higher concentrations of P4 and also possibly by pregnenolone causing activation of G_s. Therefore, the potential exists for P4 and neurosteroids to act through different mPRs and for P4 to activate an additional mPR at high concentrations in the same cell causing activation of opposite signaling pathways to either inhibit or activate protein kinase A (PKA). The marked differences in regional distribution of the mPRs in the brain, and the presence of a brain-specific mPR, also suggest that the mPRs have distinct neural functions. For example, the high expression of mPRδ and mPRβ in the hypothalamus implicates them in the control of neuroendocrine and other hypothalamic functions, whereas the high levels of mPRϵ in the pituitary gland suggest that mPRϵ may be involved, together with PR, in progestin feedback of LH secretion. Collectively, these results showing differential mPR cell and tissue expression, progestin binding specificity, and G protein activation provide a plausible explanation for their multiplicity in mammalian tissues. The presence of a complex, multifaceted system mediating nonclassical progestin actions could account for some of the pleiotropic actions of P4 and neurosteroids. The roles of the five mPRs in progestin signaling and their interactions in target cells likely represent a unique model of steroid hormone action and warrant further investigation.