Differential Regulation of Progesterone Receptor-Mediated Transcription by CDK2 and DNA-PK

Abstract

Progesterone receptor (PR) function is altered by cell signaling, but the mechanisms of kinase-specific regulation are not well defined. To examine the role of cell signaling in the regulation of PR transcriptional activity, we have utilized a previously developed mammalian-based estrogen-response element promoter array cell model and automated cell imaging and analysis platform to visualize and quantify effects of specific kinases on different mechanistic steps of PR-mediated target gene activation. For these studies, we generated stable estrogen-response element array cell lines expressing inducible chimeric PR that contains a swap of the estrogen receptor-α DNA-binding domain for the DNA-binding domain of PR. We have focused on 2 kinases important for steroid receptor activity: cyclin-dependent kinase 2 and DNA-dependent protein kinase. Treatment with either a Cdk1/2 inhibitor (NU6102) or a DNA-dependent protein kinase inhibitor (NU7441) decreased hormone-mediated chromatin decondensation and transcriptional activity. Further, we observed a quantitative reduction in the hormone-mediated recruitment of select coregulator proteins with NU6102 that is not observed with NU7441. In parallel, we determined the effect of kinase inhibition on hormone-mediated induction of primary and mature transcripts of endogenous genes in T47D breast cancer cells. Treatment with NU6102 was much more effective than NU7441, in inhibiting induction of PR target genes that exhibit a rapid increase in primary transcript expression in response to hormone. Taken together, these results indicate that the 2 kinases regulate PR transcriptional activity by distinct mechanisms.

A growing number of studies support the hypothesis that progesterone and the progesterone receptor (PR) play important roles in the development and/or progression of breast cancer. Progestins have been reported to have a carcinogenic effect in the breast in multiple animal models (reviewed in Ref. 1). Furthermore, epidemiological studies suggest that women on progestin plus estrogen hormone replacement therapy have an increased breast cancer risk compared with women on estrogen only hormone replacement therapy (2, 3). More recently, progesterone/progestins have been shown to induce expansion of progenitor cells in normal human breast and human breast cancer cell lines (4–6). Combined, these observations highlight the need to better understand how PR function is regulated in the context of breast cancer.

PR is a transcription factor that is activated in a ligand-dependent or ligand-independent manner to regulate transcription. It is expressed as 2 isoforms, PR-B and PR-A, with the latter lacking the first 164 amino acids of PR-B. In the classical, genomic model of PR action, ligands diffuse across the cell membrane and bind to PR. This binding induces a conformational change in the receptor that promotes dissociation from chaperone protein complexes, translocation to the nucleus, and dynamic binding of receptor homodimers to hormone-response elements. PR binding to DNA is accompanied by recruitment of coactivator complexes that modify chromatin to facilitate transcription (reviewed in Ref. 7). PR and coactivators are phosphoproteins with multiple known serine/threonine phosphorylation sites and their activities can be modulated through changes in phosphorylation. Several studies have implicated casein kinase 2 (8, 9), MAPK (10), cyclin-dependent kinase 2 (Cdk2) (11–13), and DNA-dependent protein kinase (DNA-PK) (14, 15) as kinases that can phosphorylate PR, whereas the p160 steroid receptor coactivator, SRC-1 was also shown to be a Cdk2 target (16). However, the mechanisms of kinase-specific regulation of PR function are not well defined.

To examine the role of cell signaling in the regulation of PR transcriptional activity, we have taken advantage of a unique systems biology-level estrogen receptor (ER)-responsive biosensor cell line (GFP-ERα:PRL-HeLa) and a custom suite of automated imaging and analysis tools developed previously (17–21). This platform is based on visualization of ER and coregulator accumulation, chromatin remodeling and transcript production and has been utilized to study mechanisms of ERα- and ERβ-mediated gene transcription. These studies include ligand specificity of ERα action (17, 18, 21, 22), functional significance of ERα domains (23), ERα association with other proteins on chromatin (19), regulation of ERα by ubiquitin ligase activity (20) and differential functional fingerprinting of ERα and ERβ in the context of endocrine disrupting chemicals (21). To adapt the system for PR, we took advantage of the well-documented modular nature of steroid receptors that enables swapping of DNA-binding domains (DBDs) between closely related steroid hormone receptors to produce functional chimeras that exhibit altered DNA binding specificity but retain cognate hormone responses (24–26). We generated a stable PRL-HeLa cell line that expresses an inducible chimeric PR-B (PR/ER) that contains a swap of the ERα DBD for the DBD of PR. In this study, the cell line was validated for the ability to faithfully exhibit mechanistic steps in progestin-induced PR-mediated gene transcription and was used to reveal insights into the role of 2 different kinases in PR function. Cdk1/2 and DNA-PK were required for efficient recruitment of RNA Pol II and production of RNA transcripts. Recruitment of selected coactivators to the array was more dependent on Cdk1/2 than on DNA-PK. Further analysis of the dependence of the kinases for hormone-dependent regulation of endogenous PR target genes indicated that Cdk1/2 may play a more important role than DNA-PK for PR-mediated primary transcriptional responses than in PR-mediated changes of mature, steady state RNA levels that could have a posttranscriptional component. These data highlight the roles of these 2 kinases in regulating distinct mechanistic actions of PR.

Materials and Methods

Materials

Cell culture reagents were obtained from Invitrogen. The Cdk1/2 inhibitor NU6102 was obtained from A.G. Scientific (N-1173). The DNA-PK inhibitor NU7441 was obtained from Selleck Chemicals (S2638). Dimethylsulfoxide was obtained from Sigma. R5020 (Promegestone) was from PerkinElmer. The PR antagonists mifepristone (RU486) and onapristone (ZK98299) were purchased from Sigma Chemical and from Schering, respectively.

Cell culture

Chimeric PR-B with a region containing the PR DBD (amino acids 565–647) swapped for a region containing the ERα DBD (amino acids 183–254) was synthesized by Epoch Life Sciences. The chimeric receptor was subcloned into a vector containing enhanced green fluorescent protein (eGFP) to obtain a GFP-tagged chimeric receptor, which was then cloned into the pHAGE-Ind-EF1a-DEST-GH vector (originally provided by Dr Jianping Jin, University of Texas Health Science Center, Houston, TX) followed by preparation of lentivirus particles. This vector places GFP-PR/ER expression under the control of a tetracycline-inducible promoter (see Supplemental Figure 1 for vector map). PRL-HeLa array cells stably expressing GFP-PR/ER (GFP-PR/ER:PRL-HeLa) were generated by transducing PRL-HeLa, enriching for cells that express GFP-PR/ER by drug selection with geneticin (G418), sorting for GFP-positive cells by flow cytometry, and then single-cell cloning to obtain a population of cells with more than 95% of cells expressing PR/ER and exhibiting hormone-dependent array formation. GFP-PR/ER:PRL-HeLa were maintained in DMEM containing 10% fetal bovine serum (FBS), hygromycin, and geneticin. Cells were plated in DMEM with 10% FBS on poly-D-lysine-coated coverslips or 384-well plates (Greiner SensoplatePlus) followed by treatment with doxycycline (200 ng/mL) for 24 hours before treatment with agonist (10nM R5020) or antagonists (10nM RU486 or 500nM ZK98299) with or without kinase inhibitor (10μM NU6102 or NU7441) pretreatment. The concentration of doxycycline treatment was empirically determined in titration assays to yield optimal array formation.

T47D cells were obtained from the American Type Culture Collection and were maintained in RPMI 1640 containing 10% FBS and 5-μg/mL insulin. For gene expression analysis, cells were plated in RPMI 1640 + 10% FBS, followed by pretreatment with 10μM NU6102 or 10μM NU7441 for 1 hour, then treatment with 10nM R5020 for the indicated time points.

Immunolabeling

For coverslips, stable GFP-PR/ER:PRL-HeLa were fixed in 4% formaldehyde (catalog number 18814, Electron microscopy grade; Polysciences, Inc) in PEM (80mM potassium PIPES, [pH 6.8], 5mM EGTA, and 2mM MgCl₂), quenched for 10 minutes with 0.1M ammonium chloride and permeabilized for 30 minutes with 0.5% Triton X-100 as described previously (17). Cells were incubated at room temperature in 5% milk in 1× Tris-buffered saline-0.02% Tween 20 (pH 7.4) for 1 hour, followed by incubation with an antibody against SRC-1 (0.5 μg/mL, 612378; BD Transduction Labs) overnight at 4°C. Cells were washed and incubated with Alexa Fluor 647 goat antimouse IgG (20 μg/mL; Molecular Probes) for 1 hour, followed by staining with 4′6-diamidino-2-phenylindole (DAPI) (1 μg/mL for 5 min) and mounting in Slowfade Gold (Molecular Probes).

For 384-well plates, GFP-PR/ER:PRL-HeLa were fixed in 4% formaldehyde in PEM buffer, quenched with 0.1M ammonium chloride and permeabilized with 0.5% Triton X-100. Cells were incubated at room temperature in 5% milk in 1× Tris-buffered saline-0.02% Tween 20 for 1 hour, followed by incubation with specific antibodies overnight at 4°C. The antibodies used were the SRC-1 antibody described above, SRC-3 (0.5 μg/mL, 611105; BD Transduction Labs), mediator 14 (MED14) (0.4 μg/mL, A301–044A; Bethyl), and Ser5-phospho RNA polymerase II (2 μg/mL, ab5401; Abcam). Cells were washed and incubated with Alexa Fluor 647 goat antimouse IgG or Alexa Fluor 647 goat antirabbit IgG (20 μg/mL; Molecular Probes). This was followed by 2 PEM washes, postfixation (10 min), and quenching (5 min) before DAPI staining (as above).

Fluorescence in situ hybridization

GFP-PR/ER:PRL-HeLa were fixed in 4% formaldehyde in ribonuclease-free phosphate-buffered saline for 15 minutes, washed in PBS, and then permeabilized with 70% ethanol in ribonuclease-free water for 1 hour at 4°C. Cells were washed in 1 mL of wash buffer (2× saline sodium citrate (SSC) with 10% formamide) followed by hybridization with custom made dsRED2 RNA probes (20) (Stellaris probes; Biosearch Technologies, Inc) in hybridization buffer (1-g dextran sulfate, 1-mL 20× saline sodium citrate (SSC) buffer and 1-mL formamide in 8 mL of nuclease-free water) for 4 hours at 37°C. After hybridization, cells were washed with wash buffer for 30 minutes at 37°C and then stained with DAPI for 10 minutes at 37°C. Cells were imaged in 2× SSC buffer.

RNAi experiments

GFP-PR/ER:PRL-HeLa were reverse-transfected using Lipofectamine RNA interefence Max reagent (Invitrogen) and 8nM of MED14 small interfering RNA (siRNA) for 48 hours, followed by treatment with doxycycline (200 ng/mL) for 24 hours, and treatment with 10nM R5020 for 1 hour. The efficiency of MED14 knockdown with the siRNA used was previously validated in GFP-ERα:PRL-HeLa (20).

Imaging and quantification

Deconvolved images were acquired using a DeltaVision Core Image Restoration Microscope (Applied Precision; GE Healthcare). Z-stacks were imaged at 0.4-μm intervals with a 40x/1.3NA oil objective and a charge-coupled device camera (CoolSnap HQ2; Photometrics) at 1 × 1 binning, and a camera gain of 4.

Automated imaging was carried out using an IC-200 image cytometer (Vala Sciences). Image acquisition was performed with a Nikon S Fluor 40x/0.90NA objective and a scientific CMOS camera. Z-stacks were imaged at 0.5-μm intervals (for a total of 3 μm) at 1 × 1 binning. Nuclear array segmentation and automated image analyses were performed using custom measurement and reporting routines in the Pipeline Pilot graphical automation software platform as previously described (20).

Real-time quantitative PCR

RNA was isolated from T47D cells treated as described above using TRIzol (Invitrogen) and reverse transcribed using amfiRivert Platinum cDNA Synthesis Master Mix from GenDEPOT. Real-time quantitative PCR was performed using SYBR Green PCR Master Mix under standard conditions on a StepOnePlus real-time PCR machine (Applied Biosystems). Primers for 18S (previously described in Ref. 27), serum/glucocorticoid-regulated kinase 1 (SGK1) mature transcript (previously described in Ref. 16), SGK1 precursor transcript (sense 5′-AGGAGGATGGGTCTGAACGA-3′, antisense 5′-AAGAAGTCTTCGCCTTCCCG-3′), FK506-binding protein 5 (FKBP5) mature transcript (previously described in Ref. 28), FKBP5 precursor transcript (sense 5′-AATGGTGAGGAAACGCCGAT-3′, antisense 5′-CCCTCTTACCTTTGCCAAGACT-3′), N-myc downstream-regulated 1 (NDRG1) mature transcript (previously described in Ref. 28), NDRG1 precursor transcript (sense 5′-GGTGAAGCCTTTGGTGGAGA-3′, antisense 5′-GGTGTGCCTGTGTGTGTCTA-3′), metallothionein 2A (MT2A) mature transcript (sense 5′-GCCCAGGGCTGCATCTG-3′, antisense 5′-TTTGTGGAAGTCGCGTTCTTT-3′), MT2A precursor transcript (previously described in Ref. 29), and Dystonin (DST/BAPG1) precursor transcript (sense 5′-TGAAGACTTAAGGGATGGACACA-3′, antisense 5′-TCAGGCAGAATTCATAGCTCAA-3′) were purchased from Sigma-Genosys. The relative mRNA expression of each transcript was determined using a standard curve and normalizing to 18S mRNA expression.

Statistical analysis

Data presented were acquired from a minimum of 3 independent experiments. Statistical significance was determined using Student's t test for single comparisons or one-way ANOVA with Tukey's post hoc test for comparisons across multiple samples. Tests were performed using GraphPad Prism 5 software, and P < .05 was considered statistically significant.

Results

One of the challenges in understanding the role of cell signaling in steroid receptor function is that steroid receptors are transcriptional modulators, and various signaling pathways may affect multiple aspects of receptor function targeting not only the receptor itself but also coregulator proteins. Thus, we wanted to establish a method to simultaneously screen for protein kinase signaling-mediated changes in multiple mechanistic steps of receptor action, including recruitment of coregulator proteins that then could be examined in a cellular context. The experimental system is a previously described ER-regulated reporter gene in HeLa cells based upon the prolactin promoter/enhancer (17) used extensively to analyze ER action. In this model, the synergy domain (eg, containing pituitary-specific positive transcription factor and estrogen-response elements [EREs]) was reiterated 52 times at the 5′ end to ensure sensitivity and microscopic visibility, and is linked to a dsRED2skl reporter. This entire cassette was then integrated approximately 100 times into the genome of HeLa cells, and reporter transcription was shown to be readily controlled by physiological levels of estrogen agonists (17, 19–22, 30). Estrogen treatment induces binding of GFP-ERα (or ERβ) to this “promoter array” resulting in a bright green intranuclear spot indicating occupancy of the ERE-rich genomic locus. Quantitative recruitment of other factors to the array can be detected using immunofluorescence with specific antibodies (a “visual chromatin immunoprecipitation” in essence) provided that the promoter array occupancy is sufficiently higher than the overall nucleoplasmic signal. Estrogen agonists have routinely led to larger, decondensed arrays indicative of chromatin remodeling, whereas ER antagonists result in smaller, condensed arrays, commensurate with increased or decreased reporter gene transcription, respectively. Combined results from multiple mechanism-linked measurements have been used to classify test compounds relative to known controls (21).

As illustrated in Figure 1, to study PR action in this system, we generated a chimeric PR by swapping a region in PR containing the DBD for a region in ERα containing the ER DBD (eg, PR/ER). The chimera was constructed as a GFP fusion protein with the tag at the amino-terminal end of PR-B. Because this microscopy-based assay is contingent upon the chimeric receptor exhibiting appropriate progestin-dependent transcriptional activity, we first sought to determine the activity of the PR/ER chimera in ERE-luciferase reporter gene assays. As a control, a GFP-tagged wild-type PR-B that mediates hormone-dependent activation of a PRE/GRE-luciferase reporter, but not of an ERE-luciferase reporter was used (Figure 1B). In contrast, the GFP-tagged PR/ER induces an ERE-luciferase reporter, but not a PRE/GRE-luciferase reporter (Figure 1B). This assay confirmed the functional switching of the PR/ER chimera. The GFP-PR/ER was also cloned into a vector containing a tetracycline-inducible promoter. We therefore generated a version of the PRL-HeLa cell line that stably expresses the tetracycline-inducible GFP-PR/ER (GFP-PR/ER:PRL-HeLa), to enable growth of the cells in the absence of PR expression and acute induction of GFP-PR/ER with doxycycline. In the absence of doxycycline, chimeric receptor protein expression was undetectable and dsRED2 reporter gene expression was at low background levels (Supplemental Figure 2). Upon doxycycline administration, chimeric GFP-PR/ER protein expression was induced and dsRED2 mRNA expression was then induced by greater than 5-fold after 2 hours of R5020 (synthetic progestin) treatment (Supplemental Figure 2). We next assessed the ability of the chimeric GFP-PR/ER to mediate responses to agonists and antagonists. In the absence of hormone (ethanol), as expected, GFP-PR/ER is primarily expressed in the cytoplasm (Figure 1C). Upon treatment with the PR agonist R5020, GFP-PR/ER translocates to the nucleus and accumulates at the promoter array (Figure 1C). GFP-PR/ER also translocates to the nucleus in the presence of the PR antagonists RU486 and ZK98299, with receptor recruitment to the array in the presence of RU486, but not ZK98299 (Figure 1C). In addition, we detected ligand-dependent recruitment of the known PR coregulator SRC-1 with promoter array occupancy only observed in the presence of R5020, but not with RU486 or ZK98299. These results are consistent with the well-characterized effects of these antagonists on PR function (31–33).

Figure 1.

Generation and characterization of stable cell lines expressing inducible chimeric PR. A, Schematic illustrating the generation of the chimeric PR by swapping a region in PR containing the DNA binding domain (amino acids 565–647) for a region of ER containing the ER DNA binding domain (amino acids 183–254) by synthesis of the chimeric sequence. B, The chimeric receptor is active on an ERE-luciferase reporter. HeLa cells were transfected with GFP-tagged wild-type PR-B or GFP-tagged chimeric receptor (PRER) and a PR-responsive GRE₂E1b-luciferase reporter (PRE-LUC) or ERE-luciferase reporter (ERE-LUC) as indicated followed by treatment with ethanol or 10nM R5020 overnight. A representative experiment from 3 independent experiments is shown. C, A stable cell line expressing inducible GFP-PRER was treated with 200 ng/mL of doxycycline for 24 hours, followed by treatment with 10nM R5020, 10nM RU486, or 500nM ZK98299 for 2 hours. Cells were then processed for immunofluorescence with an antibody against SRC-1. Images were taken on a DeltaVision Deconvolution Microscope and represent typical images from 3 independent experiments.

High throughput microscopy experiments were conducted by seeding cells in a 384-well plate, pretreating with doxycycline for 24 hours to induce expression of GFP-PR/ER, and then treating for various times ranging from a few minutes to 24 hours with R5020 alone or in combination with RU486 or ZK98299. Cells were then processed for automated image acquisition and data analysis using a custom workflow on the PipelinePilot software platform (Accelrys). This experimental approach enables a cell-by-cell quantification of nuclear translocation of binding of chimeric receptor to the array, chromatin modification, coregulator protein recruitment, and transcriptional activity at the array (measured by mRNA fluorescence in situ hybridization [FISH] of the dsRED2 reporter gene). Figure 2A illustrates examples of the nuclear and array segmentation obtained with PipelinePilot software. In the presence of R5020 there is a time-dependent increase in nuclear translocation, in the percentage of cells exhibiting promoter array occupancy by GFP-PR/ER, and in the size of the locus (large scale chromatin modification), with the peak effect occurring around 1 hour of hormone treatment (Figure 2B). We also observed a time-dependent effect on R5020-mediated transcriptional activity, with peak dsRED2 mRNA FISH intensity at the array around 1 hour and peak cytoplasmic accumulation of reporter gene mRNA at later time points (Figure 2B). Cotreatment with RU486 significantly decreased R5020-mediated GFP-PR/ER recruitment to the array (% arrays), chromatin modification (array area), and reporter gene transcriptional activity (dsRED2 FISH intensity) (Figure 2C). A similar antagonistic effect was observed with ZK98229; however, the decrease in GFP-PR/ER recruitment to the array and chromatin modification was more robust compared with RU486, consistent with the inability of ZK98299 to promote binding of GFP-PR/ER to the array (Figure 1C). Because there were a low percentage of cells with GFP-PR/ER occupancy at arrays in the presence of ZK98299, coregulator protein recruitment was only examined with RU486. Specifically, we measured the recruitment of the p160 coactivators, well-established coactivators of PR, and MED14 as this was identified as important for PR-mediated transcription in a pilot siRNA screen (data not shown). Using RU486 in combination with R5020, we observed decreased recruitment of SRC-1, SRC-3, and MED14, and decreased levels of serine 5 phosphorylated RNA Pol II, a marker of Pol II activation (Figure 2D). These data are consistent with the known mechanisms of RU486 and ZK98299 action (31–33).

Figure 2.

High-throughput microscopy analysis of the effect of the agonist R5020 and antagonists, RU486 and ZK98299, on PR transcriptional activity. A, Representative images illustrating nuclear segmentation and array segmentation with Pipeline Pilot software. Green lines indicate nuclear segmentation and red lines indicate array segmentation. B, Quantification of the percentage of cells with arrays, array area, and dsRED FISH intensity at the array and in the cytoplasm. Stable cells expressing inducible GFP-PRER were plated on a 384-well plate in full serum and then treated with 200-ng/mL doxycycline for 24 hours to induce PR expression, followed by 10nM R5020 for the indicated times. Graphs show a representative experiment from 3 independent experiments. C, Quantification of the effects of RU486 and ZK98299 on R5020-mediated percentage of cells with arrays, array area, and dsRED FISH intensity at the array. Cells were plated in full serum and treated with 200 ng/mL of doxycycline for 24 hours followed by pretreatment with ethanol, 10nM RU486, or 500nM ZK98299 for 1 hour and then 10nM R5020 for 1 hour. Graphs represent the combined data from 2 independent experiments. D, Quantification of the effect of RU486 on hormone-mediated SRC-1, SRC-3, and MED14 coregulator protein recruitment to the array, as well as phosphorylation of RNA Pol II on serine 5 (an indicator of activated RNA Pol II) at the array. Cells were pretreated with RU486, treated with R5020 as indicated in C, and then processed for immunofluorescence with antibodies against the proteins of interest. Graph represents the combined data from 3 independent experiments. Relative protein recruitment (y-axis) is the indirect immunofluorescence intensity as a result of coactivator-specific antibody binding to the coactivator bound at the array. *, P < .05; **, P < .01; ***, P < .001.

Cdk2 and DNA-PK have previously been implicated in regulating nuclear receptor transcriptional activity (11–15). Before using the Cdk1/2 inhibitor (NU6102) and the DNA-PK inhibitor (NU7441) in the new GFP-PR/ER:PRL-HeLa model, we confirmed that inhibition of both Cdk2 and DNA-PK does indeed affect wild-type PR and chimeric PR/ER transcriptional activities. Using transiently transfected luciferase reporter assays in HeLa cells, we observed that treatment with NU6102 (Figure 3A) or NU7441 (Figure 3B) substantially decreased wild-type PR and chimeric PR/ER transcriptional activity. A comprehensive dose-response experiment was conducted to determine the optimal doses of NU6102 and NU7441 that significantly decreased array area with minimal cell toxicity (data not shown). The doses chosen (10μM for each inhibitor) were in the same range as has been previously shown to completely inhibit CDK2 (34) and DNA-PK (35) activity in breast cancer cells. High throughput microscopy-based assays with GFP-PR/ER revealed that inhibition of Cdk1/2 slightly decreased R5020-mediated GFP-PR/ER recruitment to the array, but resulted in a much larger decrease of R5020-mediated chromatin modification (array area) and transcriptional activity (Figure 4A). Upon inhibition of Cdk1/2, we also observed a decreased R5020-mediated recruitment of the coregulator proteins SRC-1, SRC-3, and MED14, and decreased activation of RNA Pol II (Figure 4B). The DNA-PK inhibitor had little effect on R5020-mediated GFP-PR/ER recruitment to the array, but decreased R5020-mediated chromatin modification (array area) and transcriptional activity (Figure 4C). In contrast to the effects of the Cdk1/2 inhibitor, inhibition of DNA-PK only slightly decreased R5020-mediated recruitment of SRC-1 and activation of RNA Pol II and there was no effect on R5020-mediated recruitment of SRC-3 or MED14 (Figure 4D). The effects on coregulator protein recruitment are not due to changes in expression of the coregulator proteins as shown by immunoblot assays (Supplemental Figure 3).

Figure 3.

NU6102 and NU7441 reduce GFP-tagged PR-B and GFP-tagged chimeric PR transcriptional activity. A, NU6102 inhibits both PR-B and PRER transcriptional activity. HeLa cells were transfected with GFP-PR-B, GFP-PRER, GRE₂E1b-LUC, and ERE-LUC as indicated. Cells were pretreated with 10μM NU6102 for 1 hour followed by 10nM R5020 overnight. Graphs are indicative of 3 independent experiments. B, NU7441 inhibits both PR-B and chimeric PR-B transcriptional activity. HeLa cells were transfected with GFP-PR-B, GFP-PRER, GRE₂E1b -LUC, and ERE-LUC as indicated. Cells were pretreated with 10 μM NU7441 for 1 hour followed by 10nM R5020 overnight. Graphs are indicative of 3 independent experiments. ***, P < .001.

Figure 4.

Kinase-specific regulation of PR transcriptional activity by Cdk2 and DNA-PK. A, Quantification of the effect of Cdk2 inhibition on hormone-mediated PR recruitment, array area, and dsRED FISH intensity at the array. Cells were plated in full serum, treated with 200 ng/mL of doxycycline for 24 hours, then pretreated with 10μM NU6102 for 1 hour, followed by treatment with 10nM R5020 for 1 hour. Graphs indicate a representative experiment from 2 independent experiments. B, Quantification of the effect of NU6102 on hormone-mediated SRC-1, SRC-3, and MED14 coregulator protein recruitment to the array and phosphorylation of RNA Pol II on serine 5 (an indicator of activated RNA Pol II) at the array. Cells were pretreated with NU6102, treated with R5020 as indicated in A, and then processed for immunofluorescence with antibodies against the proteins of interest. Graph represents the combined data from 3 independent experiments. C, Quantification of the effect of DNA-PK inhibition on hormone-mediated PR recruitment, array area, and dsRED FISH intensity at the array. Cells were plated in full serum, treated with 200 ng/mL of doxycycline for 24 hours, then pretreated with 10μM NU7441 for 1 hour, followed by treatment with 10nM R5020 for 1 hour. Graphs indicate a representative experiment from 2 independent experiments. D, Quantification of the effect of NU7441 on hormone-mediated SRC-1, SRC-3, and MED14 coregulator protein recruitment to the array and phosphorylation of RNA Pol II on serine 5 (an indicator of activated RNA Pol II) at the array. Cells were pretreated with NU7441, treated with R5020 as indicated in A, and then processed for immunofluorescence with antibodies against the proteins of interest. Graph represents the combined data from 3 independent experiments. *, P < .05; **, P < .01; ***, P < .001; ****, P < .0001.

As mentioned previously, R5020-mediated recruitment of MED14 to target DNA was decreased upon treatment with RU486. MED14 is a component of the preinitiation complex and plays a role in nuclear receptor transcription, but little is known about its role in PR-mediated transcription. To assess the role of MED14 in PR action, we reduced MED14 expression using siRNA. Decreased expression of MED14 correlated with decreased R5020-mediated chromatin modification (array area) and GFP-PR/ER transcriptional activity (Figure 5).

Figure 5.

MED14 is required for optimal PR transcriptional activity. Quantification of the effect of MED14 depletion on PR recruitment to the array, array area, and dsRED FISH intensity at the array. Cells were plated in full serum, treated with 200 ng/mL of doxycycline for 24 hours, and then treated with 10nM R5020 for 1 hour after 48-hour depletion of MED14. Graphs represent the combined data from 3 independent experiments. ****, P < .0001.

We next sought to extend our PRL-HeLa model results to determine the effects of kinase inhibition in the T47D breast cancer cell that expresses endogenous PR. The genes selected for this analysis have been previously described as direct PR targets (36–38) and as responsive to cyclin A depletion (NDRG1, FKBP5, and SGK1) or a broader cyclin dependent kinase inhibitor, roscovitine (MT2A) (39, 40). Cyclin A/Cdk2 has been shown previously to be important for optimal induction of PR-mediated transcription (16, 39, 40). We observed a gene-specific effect of inhibition of Cdk1/2 on endogenous PR target gene expression. For example, inhibition of R5020-mediated induction of NDRG1 and MT2A occurred at both 6 hours and 24 hours, whereas inhibition of induction of FKBP5 was observed at 24 hours but not at 6 hours, and there was little effect on induction of SGK1 at either 6 or 24 hours (Figure 6A). We also observed a gene-specific effect of inhibition of DNA-PK with significant decreases in R5020-mediated induction of NDRG1, MT2A, and SGK1 at 24 hours of hormone with no effect at 6 hours and no effect on induction of FKBP5 at either time of hormone treatment (Figure 6B).

Figure 6.

Kinase-specific effect on PR-mediated transcription of endogenous target genes in T47D. A, Effect of Cdk2 inhibition on expression of endogenous PR target genes. T47D cells were plated in full serum, pretreated with 10μM NU6102 for 1 hour, and then treated with 10nM R5020 for the indicated time. RNA was extracted and reverse transcribed followed by real-time qPCR. Graphs indicate representative data for 3 independent experiments. B, Effect of DNA-PK inhibition on expression of endogenous PR target genes. T47D cells were plated in full serum, pretreated with 10μM NU7441 for 1 hour and then treated with 10nM R5020 for the indicated time. RNA was extracted and reverse transcribed followed by real-time qPCR. Graphs indicate representative data for 3 independent experiments. *, P < .05; **, P < .01; ***, P < .001.

The effects observed on NDRG1 and MT2A gene expression with inhibition of Cdk2 at 6 hours of hormone treatment (Figure 6A) suggests that Cdk2 may play a role in PR-mediated activation of primary gene transcription. To examine this further, primers were designed to measure the expression of preprocessed primary RNA transcripts or to measure the steady state levels of mature RNA (Figure 7A). We observed robust induction of primary RNA transcript expression of NDRG1 and MT2A, but not FKBP5 or SGK1, with short-term (30 min) R5020 treatment (Supplemental Figure 4). Furthermore, the Cdk1/2 inhibitor NU6102 almost completely abrogated R5020-mediated induction of NDRG1 and MT2A primary RNA transcript expression (Figure 7B). On the other hand, the DNA-PK inhibitor NU7441 modestly inhibited R5020-mediated induction primary RNA transcripts of NDRG1 and had no effect on MT2A (Figure 7C). To examine the influence of these kinases on PR-mediated gene repression we examined the effects of NU7441 and NU6102 on a previously defined target gene that is actively repressed by progestins (40). There was no effect of Cdk2 or DNA-PK inhibition on PR-mediated progestin dependent reduced expression of Dystonin/BPAG1 (Figure 7D).

Figure 7.

Kinase-specific effect on PR-mediated transcription of precursor target genes in T47D. A, Schematic of primers designed to detect either the preprocessed primary RNA transcript or the mature, steady state mRNA transcript using real-time qPCR. B, Effect of Cdk2 inhibition on expression of precursor genes. T47D cells were plated in full serum, pretreated with 10μM NU6102 for 1 hour and then treated with 10nM R5020 for 30 minutes. RNA was extracted and reverse transcribed followed by real-time qPCR. Graphs indicate representative data for 3 independent experiments. C, Effect of DNA-PK inhibition on expression of precursor genes. T47D cells were plated in full serum, pretreated with 10μM NU7441 for 1 hour, and then treated with 10nM R5020 for 30 minutes. RNA was extracted and reverse transcribed followed by real-time qPCR. Graphs indicate representative data for 3 independent experiments. **, P < .01. D, Effect of Cdk2 or DNA-PK inhibitors on progestin-dependent repression of DSTP1/BPAG1. T47D cells were plated in full serum, pretreated with 10μM NU6102 or NU7441 for 1 hour, and then treated with 10nM R5020 for 6 or 24 hours. RNA was extracted and reverse transcribed followed by real-time qPCR. Graphs indicate representative data for 3 independent experiments. *, P < .05.

Figure 8 summarizes the kinase-specific effects on PR-mediated transcription. Inhibition of Cdk1/2 or DNA-PK decreased overall transcriptional activity with a concomitant decrease in recruited phosphorylated (activated) RNA Pol II and an inhibition of chromatin modification. Neither Cdk1/2 nor DNA-PK inhibitors affected nuclear translocation or array binding of the chimeric receptor. The major difference between Cdk2 and DNA-PK is that inhibition of Cdk1/2 decreased the recruitment of the p160 coregulator proteins (SRC-1, SRC-2, and SRC-3) and a member of the mediator complex, MED14, whereas inhibition of DNA-PK had little to no effect on the recruitment of these coregulator proteins. However, coregulators are phosphorylated on multiple sites and we cannot exclude reduced Cdk or DNA-PK-mediated phosphorylation of these coregulators as a mechanism for inhibiting PR-mediated transcription.

Figure 8.

Model of Cdk2-mediated and DNA-PK-mediated regulation of PR transcriptional activity. Cdk2 inhibition decreases chromatin modification, p160 coregulator protein recruitment, MED14 recruitment, activation of RNA Pol II, and transcriptional activity. DNA-PK inhibition decreases chromatin modification, activation of RNA Pol II, and transcriptional activity but has little to no effect on p160 coregulator protein recruitment.

Discussion

We combined a high throughput cell imaging method originally developed for ER studies with examination of endogenous PR and endogenous target genes to elucidate the steps in PR action regulated by 2 kinases. Extending previous high throughput single cell analysis work on ER functions, the PR/ER chimera model developed here enabled the direct visualization and quantification of multiple mechanistic steps involved in PR transcription function. The new cell line allowed simultaneous quantitative cell-by-cell analysis of GPF-PR/ER translocation, binding to the ERE-rich prolactin (PRL) promoter array, localized chromatin modification at the array, coregulatory protein recruitment, and reporter gene transcriptional activity (measured by mRNA FISH). As with the original ER PRL arrayed cell line, the chimeric PR/ER model recapitulates the mechanism of actions of PR agonists and antagonists using conventional biochemical and molecular approaches. There are multiple advantages of using this system to interrogate mechanisms regulating receptor transcriptional activity. First, it provides a quantitative high throughput assay to screen for recruitment of proteins to the GFP-PR/ER-occupied array transcription locus provided that the enriched accumulation on the array is above the general nucleoplasmic background. Second, this approach can be used in siRNA screens for candidates that regulate receptor activity as shown in (20). In fact, MED14 was identified in a pilot screen of coregulators. One of the challenges in studying the role of cell signaling in receptor activity is the potential for secondary effects such as cell cycle synchronization with long-term incubation. As such, another strength of this experimental system is that short-term effects of inhibitors on PR activation can readily be detected in 60 minutes, without significant changes to cell cycle. The dsRED2 probe covers multiple portions of the transcript and the initial localization at the array provides a sensitive means to measure early changes in transcription; indeed, early time course studies revealed some GFP-PR/ER cells respond within minutes of treatment.

With the GFP-PR/ER:PRL-HeLa cell line, we observed that chemical inhibition of Cdk1/2 resulted in a significant decrease in array levels of phosphorylated RNA polymerase II and chimeric PR/ER transcriptional activity. This is consistent with previous reports from our lab that showed Cdk2 is required for hormone-dependent activation of some PR target genes (16, 39, 40). Although cyclin A/Cdk2 can directly phosphorylate PR, this does not appear to be the main mechanism for regulation of PR transcriptional activity by Cdk2 (39). Previous studies using a general Cdk inhibitor suggested that Cdk inhibition reduced PR binding to SRC-1 (39), and cyclin A/Cdk2 has been shown to phosphorylate SRC-1 (16, 39). In this study, inhibition of Cdk1/2 activity resulted in decreased recruitment of SRC-1, SRC-3, and MED14 to the hormone-response element array. Whether this is due to Cdk-dependent modifications of all 3 proteins or through another mechanism remains to be determined. The potential for Cdk2 to phosphorylate and regulate SRC-3 is significant due to its well-known role in breast tumorigenesis (reviewed in Ref. 41). To our knowledge, no Cdk2-mediated phosphorylation of MED14 has been reported and little is known about the regulation of PR transcriptional activity by MED14. One study showed that nicotinamide decreases PR-mediated transcription of a reporter gene presumably due to decreased recruitment of components of the preinitiation complex, including MED14, to DNA (42). We observed that knockdown of MED14 results in significantly decreased chromatin modification (eg, formation of smaller, less decondensed arrays) and chimeric PR transcriptional activity, supporting an important role for MED14 in regulating PR target gene transcription. It is possible that MED14 may provide a mechanism of promoter selectivity as has been shown for glucocorticoid receptor (43), although this was not directly tested in our experiments.

In addition to Cdk2-mediated phosphorylation of coregulatory proteins, Cdk2 has also been shown to phosphorylate histone H1, facilitating chromatin modification and PR transcriptional activity (44). A recent study reported that Cdk2 also phosphorylates and activates poly ADP-ribose polymerase, which further facilitates displacement of histone H1, and the recruitment of chromatin remodelers and other coregulatory proteins (45). Although we observed that inhibition of Cdk1/2 significantly decreased agonist-induced array decondensation, we were unable to reliably assess histone H1 levels specifically at the array due to its extensive distribution throughout the nucleoplasm; as such, we cannot determine from the present studies whether H1 plays a role in Cdk2-mediated chromatin modification.

To assess changes in transcription of endogenous genes, we employed primers that measured preprocessed primary RNA transcripts as well as primers that measured changes in mature steady state mRNAs. We found that PR-dependent hormone-induced mRNA synthesis of NDRG1 and MT2A was rapid (within 30 min of R5020 treatment), but that induction of FKBP5 and SGK1 was much slower and that both primary transcripts and total mRNA levels were increased by hormone (Supplemental Figure 4). Thus, the primary role of PR in regulating FKBP5 and SGK1 either does not involve enhancement of initiation or there is an additional rate-limiting factor. Interestingly, whereas Cdk1/2 inhibition profoundly inhibited the induction of all of the highly regulated artificial promoters (ERE-luciferase and the array), the effects on endogenous genes, which are also regulated by other transcription factors, were more variable. Cdk1/2 inhibition strongly reduced hormone-mediated induction of the 2 endogenous genes whose regulation appears to be at the level of initiation of transcription (NDRG1 and MT2A), but had more modest effects on the other 2 genes (FKBP5 and SGK1). Inhibition of neither Cdk1/2 nor DNA-PK had an effect on PR-mediated repression of DST1/BPAG1. This is consistent with previous studies utilizing long-term depletion of cyclin A2 (40), and suggests that activity of these kinases may not be necessary for PR-mediated repression (at least of DST1/BPAG1). Additional studies with multiple repressed PR targets are warranted.

Although inhibition of DNA-PK significantly decreased activation of RNA Pol II and chimeric PR transcriptional activity similar to the effects observed upon inhibition of Cdk2, the measurements of coregulator recruitment and pattern of regulation of endogenous target genes show that it plays a role distinct from Cdk2. Previous studies have shown that DNA-PK can phosphorylate PR in a DNA-dependent manner (14, 15), and it is possible that DNA-PK might modulate the recruitment of select coregulator proteins. For example, DNA-PK mediates phosphorylation of ERα (46) that dynamically alters the components of the coregulator-ERα-DNA complex (47). The current study suggests that actions other than SRC-1, SRC-3, and MED14 recruitment are important for DNA-PK effects on mediated PR transcription because inhibition of DNA-PK had little to no effect on recruitment of these select coregulatory proteins in the PRL-HeLa cell line. Mass spectrometry studies are currently underway to identify PR coregulatory proteins that are DNA-PK substrates. It will be interesting to compare and contrast this list with the reported DNA-PK targets reported for ERα coregulatory proteins (47).

The inhibitor studies with GFP-PR/ER:PRL-HeLa cells showed a kinase-specific effect on PR-mediated transcription of the array driven reporter gene. Importantly this was also observed with endogenous PR target genes in T47D cells, where inhibition of Cdk2 had a much greater effect than inhibition of DNA-PK, on hormone induced gene expression at the level of synthesis of primary transcripts. This suggests that Cdk2 is important at an early step of transcriptional activation. In these studies, we did not determine whether there is a differential effect of the Cdk1/2 and DNA-PK inhibitors on SRC-1, SRC-3, or MED14 recruitment to endogenous target genes. It is also not known whether these coregulatory proteins play a role in PR-mediated stimulation of primary target gene transcripts. Although DNA-PK inhibition substantially decreased chromatin modification and PR transcriptional activity at the array, a smaller effect was observed on PR-mediated transcription of endogenous genes in T47D cells. This may be due to the fact that the arrays are minimal hormone-response element gene constructs designed, in part, to be as receptor dependent as possible. Under these conditions, PR alone must recruit the components that modify chromatin and induce transcription. On the other hand, the endogenous PR target genes examined exhibit a strong fold induction by hormone, but also have a substantial level of transcription in the absence of hormone, presumably through actions of other transcription factors. Thus, some of the modifications that may depend on PR-mediated kinase recruitment may have already occurred on actively transcribed promoters of these particular endogenous targets. A role for DNA-PK in regulating the initial transcription rate of at least a subset of genes cannot be ruled out, as we have observed a modest reduction in expression of NDRG1 primary transcripts upon inhibition of DNA-PK. This reduction, however, was insufficient to reduce overall transcript levels at 6 hours although there is a reduction in overall transcript levels at 24 hours.

It is clear that PR plays a role in the development and/or progression of breast cancer and understanding the crosstalk between PR and cell signaling pathways may have clinical relevance. Several cell signaling pathways have been identified as putative therapeutic targets for breast cancer (reviewed in Ref. 48). Furthermore, recent studies suggest that measuring activation of signaling pathways in tumors before and after therapeutic intervention can aid in clinical management of the disease (reviewed in Ref. 49). In conclusion, we have developed a modified PRL/ERE-HeLa cell line and high throughput microscopy-based experimental system that provides visualization and quantitation of the effect of kinases on PR functions. With the proven synergy and versatility of the cell and imaging platforms described in this work, we are now poised to screen and compare additional cell signaling pathways that regulate activities and putative therapeutic targets of PR.

Additional material

Supplementary data supplied by authors.