The Selective Progesterone Receptor Modulator Ulipristal Acetate Inhibits the Activity of the Glucocorticoid Receptor

Abstract

Context

The selective progesterone modulator ulipristal acetate (ulipristal) offers a much-needed therapeutic option for the clinical management of uterine fibroids. Although ulipristal initially passed safety evaluations in Europe, postmarketing analysis identified cases of hepatic injury and failure, leading to restrictions on the long-term use of ulipristal. One of the factors potentially contributing to significant side effects with the selective progesterone modulators is cross-reactivity with other steroid receptors.

Objective

To determine whether ulipristal can alter the activity of the endogenous glucocorticoid receptor (GR) in relevant cell types.

Design

Immortalized human uterine fibroid cells (UtLM) and hepatocytes (HepG2) were treated with the synthetic glucocorticoid dexamethasone and/or ulipristal. Primary uterine fibroid tissue was isolated from patients undergoing elective gynecological surgery and treated ex vivo with dexamethasone and/or ulipristal. In vivo ulipristal exposure was performed in C57Bl/6 mice to measure the effect on basal gene expression in target tissues throughout the body.

Results

Dexamethasone induced the expression of established glucocorticoid-target genes period 1 (PER1), FK506 binding protein 51 (FKBP5), and glucocorticoid-induced leucine zipper (GILZ) in UtLM and HepG2 cells, whereas cotreatment with ulipristal blocked the transcriptional response to glucocorticoids in a dose-dependent manner. Ulipristal inhibited glucocorticoid-mediated phosphorylation, nuclear translocation, and DNA interactions of GR. Glucocorticoid stimulation of PER1, FKBP5, and GILZ was abolished by cotreatment with ulipristal in primary uterine fibroid tissue. The expression of glucocorticoid-responsive genes was decreased in the lung, liver, and uterus of mice exposed to 2 mg/kg ulipristal. Interestingly, transcript levels of Fkbp5 and Gilz were increased in the hippocampus and pituitary.

Conclusions

These studies demonstrate that ulipristal inhibits endogenous glucocorticoid signaling in human fibroid and liver cells, which is an important consideration for its use as a long-term therapeutic agent.

Uterine leiomyomas, fibroids arising from the smooth muscle layer of the uterus, are the most common tumors found in women of reproductive age, where it is estimated that 70% to 80% of women will have 1 or more uterine fibroid tumor by the age of 50 (1). Although these tumors are considered benign, uterine fibroids cause significant gynecologic and reproductive dysfunction. Women with uterine fibroids experience heavy uterine bleeding, anemia, voiding problems, pelvic pain, preterm birth, postpartum comorbidities, and rare transformation to malignant leiomyosarcoma (2–4). The associated symptoms compromise a woman’s physical, social, and emotional well-being, resulting in a lower quality of life (4, 5). Despite their prevalence, severity, and health care burden, there are limited noninvasive treatment options for women with uterine fibroids, and none are US Food and Drug Administration–approved for long-term use. As such, the only definitive treatment that can eliminate uterine fibroids and prevent recurrence is hysterectomy. Although hysterectomy remains the first-line treatment, the development of medical therapeutics would eliminate the need for surgery and preserve the uterus of women afflicted with uterine fibroids who want to preserve their fertility (6).

The pathogenesis of uterine fibroids is driven by the female sex steroids, and accordingly, the cells composing the fibroid mass express high levels of the receptors for estrogen and progesterone (PR) (7–12). This hallmark of uterine fibroids has been targeted for the development of nonsurgical medical treatments. Gonadotropin-releasing hormone agonists were 1 of the first therapies shown to successfully decrease uterine fibroid size by inducing a hypoestrogenic state (13, 14). However, this pseudomenopausal condition resulted in a wide range of significant side effects, which limited the use of gonadotropin-releasing hormone agonists to short-term durations (15). Newer treatments have focused on synthetic steroid ligands that possess tissue-selective activities to achieve efficacy without the limiting hypoestrogenic state.

The selective progesterone receptor modulators (SPRM) were developed to specifically target the actions of PR in target tissue. As such, recent studies have focused on the therapeutic targeting of PR to exploit the hormone-responsive nature of uterine fibroids (14). Several SPRMs, including mifepristone (RU-486), asoprisnil, telapristone, vilaprisan, and ulipristal acetate (ulipristal [Uli]), have shown promise in improving uterine fibroid size, associated symptoms, and reported quality-of-life scores (14, 16–22). Among the most promising, ulipristal was investigated in 6 phase 3 trials in Europe and the United States and subsequently approved for the long-term treatment of uterine fibroids in Europe and Canada (23–25). However, safety concerns related to pathology in the endometrium and several reported cases of liver failure have halted the long-term use of ulipristal and led the US Food and Drug Administration to reject the new drug application for ulipristal in the United States (26, 27). Because the liver does not express appreciable levels of PR, it becomes critical to understand the etiology of liver damage associated with the long-term use of ulipristal, which may reflect drug metabolism toxicity or the modulation of other signaling pathways (28, 29).

One of the limiting factors in the development of SPRMs has been their cross-reactivity with other steroid receptors. Interestingly, the SPRM RU-486 was originally developed as part of a program to identify antiglucocorticoids but also possessed significant antiprogestin activity (30). RU-486 was quickly recognized for its clinical utility in PR-regulated functions. However, RU-486 displays high affinity for the closely related glucocorticoid receptor (GR) and demonstrates potent antiglucocorticoid activities in the liver (31–33). The severe side effects associated with disruption to physiological glucocorticoid signaling (including hepatotoxicity) is a major limiting factor for the long-term use of RU-486 (34–36). Ulipristal was found to possess a 50% lower binding affinity for isolated rabbit GR compared with RU-486 and decreased antagonist activity in a HepG2 reporter assay, inferring improved selectivity and greater promise as a therapeutic candidate (37). However, others reported that RU-486 and ulipristal had relatively similar binding affinities using GR isolated from human breast cancer cells (38). Interestingly, these authors demonstrated that although ulipristal was less efficient than RU-486 in antagonizing GR transcriptional activity using a reporter construct, ulipristal could inhibit the nuclear translocation of GR necessary for receptor activity. Therefore, ulipristal may act as a competitive GR antagonist, by competing for binding and preventing the movement of GR into the nucleus. Importantly, the antiglucocorticoid activity of ulipristal has not been evaluated on endogenous GR in human uterine fibroid and liver cells.

In this study, we tested the hypothesis that ulipristal potently inhibits endogenous GR action in human uterine fibroid and liver cells. Here, we show that ulipristal restricts the cellular response to glucocorticoids by blocking GR phosphorylation, preventing nuclear translocation, and significantly reducing subsequent GR–DNA interactions. In primary human fibroid tissue, ulipristal completely abolished the upregulation of classic glucocorticoid-responsive genes. In vivo exposure demonstrated tissue specificity in the actions of ulipristal on GR, including GR antagonist activity in the liver and uterus and agonist activity in the brain. Our findings indicate that ulipristal possesses antiglucocorticoid activity comparable to RU-486 and that treatment with ulipristal would be expected to disrupt endogenous glucocorticoid signaling.

Materials and Methods

Reagents

Fetal bovine serum (FBS) was purchased from Sigma-Aldrich (F2442; St. Louis, MO). Charcoal-dextran treated (stripped) FBS was purchased from Gemini Bio-Products (100-119; Sacramento, CA). Minimum essential media (MEM; 11095-080) and DMEM (10-013-CV) were purchased from Invitrogen (ThermoFisher Scientific; Waltham, MA). MEM amino acids (11130-051), MEM vitamin solution (11120-052), and MEM nonessential amino acids (11140-050) were purchased from Gibco (ThermoFisher Scientific). NR3C1 (L-003424-00-0005) and nontargeting control (NTC; D-001810-10-05) ON-TARGET Plus SMART pool small interfering RNA (siRNA) were purchased from Dharmacon (GE Healthcare, Chicago, IL). Predesigned TaqMan quantitative RT-PCR (qRT-PCR) primer-probe sets (39) were purchased from Applied Biosystem (ThermoFisher Scientific). Dexamethasone (Dex; 1, 4-pregnadien-9α-fluoro-16α-methyl-11β, 17, 21-triol-3, 20-dione; ≥98% by thin-layer chromatography [TLC]), cortisol (4-pregnen-11β, 17, 21-triol-3, 20-dione; 98% by TLC), and mifepristone (RU-486; 11β-[4-(dimethylamino)phenyl]-17β-hydroxy-17-(1-propynyl)-estra-4,9-dien-3-one; ≥97% by TLC) were purchased from Steraloids (Newport, RI). Norethisterone acetate (NOAC; >98% by HPLC) was purchased from TCI America (Portland, OR). Uli (8S,11R,13S,14S,17R)-17-acetyl-11-[4-(dimethylamino)phenyl]-13-methyl-3-oxo-1,2,6,7,8,11,12,14,15,16-decahydrocyclopenta[a]phenanthren-17-yl] acetate; ≥98% by HPLC) was purchased from Sigma-Aldrich.

Cell culture

All cells were maintained in a standard tissue culture incubator at 37°C in 5% CO₂. HEPG2 cells were kindly provided by Dr Yingqun Huang and grown in DMEM supplemented with 1.0 nM sodium pyruvate, 10% heat-inactivated FBS, 100 units/mL penicillin, 100 µg/mL streptomycin, and 2 mm glutamine. One day before experiments, media were changed to phenol red-free media supplemented with 10% charcoal dextran-stripped, heat-inactivated FBS, 1.0 nM sodium pyruvate, 100 units/mL penicillin, 100 µg/mL streptomycin, and 2 mm glutamine. Immortalized human uterine fibroid cells (UtLM) were kindly provided by Dr Ayman Al-Hendy and grown in MEM supplemented with 10% heat-inactivated FBS, 2 mm glutamine, 1X MEM amino acids, 1X MEM vitamin solution, and 0.10 mM MEM nonessential amino acids. The pH of the UtLM growth media was adjusted to 7.4. One day before experiments, media were changed to phenol-free media supplemented with 10% charcoal dextran-stripped, heat-inactivated FBS, 2 mm glutamine, 1X MEM amino acids, 1X MEM vitamin solution, and 0.10 mM MEM nonessential amino acids. Cell lines were authenticated by short-tandem repeat analysis using the DNA Analysis Facility at Yale University.

RNA interference

Cells plated at approximately 70% confluency were transfected with 50 nM siRNA against NR3C1 or NTC using DharmaFECT1 transfection reagent (T-2001-02; ThermoFisher Scientific) according to the manufacturer’s protocol. The following day, cells were split for experimental endpoint.

Human samples

Deidentified uterine fibroid tissue samples were obtained from 4 reproductive-age women undergoing elective gynecological surgery for benign conditions at Yale New Haven Hospital (New Haven, CT). Women were excluded if they received hormone therapy in the 3 months before surgery. The age range of included women was 33 to 40 years. The use of this deidentified tissue was approved by the Yale University Human Investigations Committee. Fibroid tissue was collected in 1X PBS containing 5% penicillin/streptomycin and transported directly to the laboratory. Tissue was rinsed in 1X PBS and diced into 30-mg explants. A portion of the dissected tissue from each fibroid sample was then fixed in 10% buffered formalin. The remaining 30-mg explants were plated on top of 6.5-mm transwell inserts (Cat. # 3422; Corning, Corning, NY) containing polycarbonate membranes with an 8-μm pore size. Explants were incubated at 37°C overnight with phenol red-free DMEM supplemented with 10% charcoal dextran-stripped, heat-inactivated FBS, 1.0 nM sodium pyruvate, 100 units/mL penicillin, 100 µg/mL streptomycin, and 2 mm glutamine placed above and below the insert. The following day, explants were treated for 6 hours with vehicle (1X PBS), 100 nM Dex, and 1 µM Uli, or 1 µM Uli and 100 nM Dex. For the combination treatment, Uli was added 30 minutes before addition of Dex. After treatment, explants were transferred to RNAlater RNA Stabilization Reagent (Qiagen, Valencia, CA) and stored at 4°C until RNA extraction was performed.

Animal experiment study design

All experiments were conducted with the approval of the Institutional Animal Care and Use Committee at Yale University and followed the National Institute of Health Guide for the Care and Use of Laboratory Animals. Adult female C57Bl/6 mice were maintained in a controlled 12:12 hour light-dark cycle (lights on 0700-1900 hours) and provided with ad libitum access to water and food. Mice in the diestrus phase of the estrous cycle were randomly assigned to receive either vehicle (corn oil) or Uli (2 mg/kg dissolved in dimethyl sulfoxide and diluted in corn oil) by intraperitoneal injection. The dose of ulipristal was selected based on the PGL4001 Efficacy Assessment in Reduction of Symptoms due to Uterine Leiomyomata (PEARL) studies (5- to 10-mg dose) and scaled appropriately for mice (40). Tissue from the lung, liver, spleen, uterus, hippocampus, and pituitary were harvested 4 hours after injection for RNA extraction using RNAlater RNA Stabilization Reagent (Qiagen) and stored at 4°C until RNA extraction was performed.

RNA isolation

Mouse tissues and ex vivo human samples were homogenized using the PRO Scientific Bio-Gen PRO200 Homogenizer followed by passage through a QIAshredder column (79654; Qiagen). Total RNA was harvested from tissue samples and cultured cells using the Qiagen RNeasy mini kit (74106; Qiagen), with the deoxyribonuclease (DNase) treatment performed on the column per the manufacturer’s instructions. The NanoDrop One Spectrophotometer (ThermoFisher Scientific) was used to assess the RNA purity and yield based on the absorbance ratios of 260 to 280 nM and 260 to 230 nM.

Quantitative RT-PCR

From 100 ng of input RNA, complementary DNA was synthesized using the One-Step RT-PCR Universal Master Mix reagent (172-5141; ThermoFisher Scientific). qRT-PCR was performed using the CFX Connect or CFX384 thermocycler (BioRad) in a reaction volume of 10 μL. The thermocycling parameters for each reaction were 48°C for 30 minutes, 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 60 seconds. Each of the predesigned primer-probe sets (Table S1) was analyzed in technical duplicates using a standard curve and compared with the reference gene peptidylprolyl isomerase B (PPIB; Ppib). When an insufficient quantity of RNA was available from ex vivo human samples to create a standard curve, the delta-delta Ct method was used to determine relative transcript levels from the gene of interest and the reference gene PPIB.

Western blotting

Treated HEPG2 and UtLM cells were lysed in Tris-glycine SDS sample buffer (LC2676; Life Technologies) supplemented with 2-mercaptoethanol. Nuclear/cytoplasmic fractionation was performed using the NE-PER Nuclear and Cytoplasmic Extraction Reagents (78833; ThermoFisher Scientific) per the manufacturer’s protocol with an approximate packed cell volume of 20 µL. Protein was isolated and quantified using the Pierce 660 nM Protein Assay (1861426; ThermoFisher Scientific) and the NanoDrop One Spectrophotometer. Equal quantities of protein were separated by SDS-PAGE and transferred to a nitrocellulose membrane using the Trans Blot Turbo System (BioRad). Membranes were stained with a Ponceau solution to visualize equivalent loading. Membranes were blocked with 7.5% nonfat milk or 2% BSA in Tris-buffered saline and then probed overnight with specific antibodies (39): rabbit monoclonal antiphosphorylated GR (pGR) s211 antibodies (4161S; Cell Signaling; 1:1000), rabbit monoclonal anti-pGR s226 antibodies (97285S; Cell Signaling; 1:1000), rabbit monoclonal anti-GR antibodies (3660S; Cell Signaling; 1:1000), rabbit polyclonal anti-PR antibodies (A301-200A-M; Bethyl Laboratories, Montgomery, TX; 1:1000), or mouse monoclonal anti-β-actin antibodies (3018859; Millipore; 1:3000) in 5% milk in Tris-buffered saline-Tween. The next day, membranes were washed and incubated with goat anti-rabbit IRDye 800-conjugated secondary antibody or goat anti-mouse IRDye 680-conjugated secondary antibody (LI-COR) for 1 hour at room temperature. The Odyssey LI-COR imaging system was used to evaluate protein expression. Protein levels were normalized to the reference protein β-actin and expressed relative to control samples.

Immunohistochemistry

Human uterine fibroid sections from paraffin-embedded tissue (5 µm) were deparaffinized in xylenes and rehydrated in a graded alcohol series. Antigen retrieval was conducted using a 10 mM citrate buffer, pH 6.0, and endogenous peroxidases were quenched with 2% hydrogen peroxide in PBS. Sections were then blocked for 1 hour with 10% normal goat serum (NGS) in PBS and incubated overnight with antibodies against GR (3660S; Cell Signaling; 1:500). The following day, sections were washed with PBS containing 0.1% Tween‐20 and incubated for 1 hour with biotinylated goat anti-rabbit IgG antibody (Vector Laboratories, Burlingame, CA). The antigen–antibody complex was visualized using VECTASTAIN ABC Kit (PK-6100; Vector Laboratories) and DAB Peroxidase Substrate Kit (SK-4100; Vector Laboratories). Sections were counterstained with Harris Hematoxylin (s212; Poly Scientific R&D Corp, Bay Shore, NY), dehydrated, and mounted for imaging. Images were captured using the Revolve 3 microscope at ×20 (Echo, San Diego, CA).

Immunofluorescence

HepG2 and UtLM cells were treated for 1 hour with vehicle (1X PBS), 100 nM Dex, and 1 µM Uli, or 1 µM Uli and 100 nM Dex. For the combination treatment, Uli was added 30 minutes before addition of Dex. Treated cells were washed twice for 5 minutes in cold 1X PBS and fixed for 20 minutes in 4% paraformaldehyde. Cells were then washed 3 times for 5 minutes in 1X PBS and permeabilized for 30 minutes in 2% BSA and 0.15% Triton X-100 in 1X PBS. Cells were probed overnight with antibodies against GR (3660S; Cell Signaling; 1:1000) in 2% BSA with 0.15% Triton X-100. The following day, cells were washed 5 times in 1X PBS, blocked for 30 minutes in 5% NGS in 1X PBS, and incubated for 1 hour in anti-rabbit Alexa Fluor 488 conjugated secondary antibody (1:500) in 5% NGS. Cells were then washed 5 times in 1X PBS and mounted using the ProLong Gold antifade reagent with DAPI (Invitrogen). Images were captured using a ×63 objective on the Zeiss 880 confocal microscope and processed with the Zen 2.1 software.

Chromatin immunoprecipitation

HepG2 and UtLM cells were treated for 1.5 hours with vehicle, 100 nM Dex, or 1 µM Uli and 100 nM Dex. For the combination treatment, Uli was added 30 minutes before addition of Dex. Treated cells were fixed with cold 1% paraformaldehyde diluted in 1X PBS for 10 minutes at room temperature. Cells were washed in ice-cold PBS, and cross-linking was stopped with 125 nM glycine. Cells were pelleted by centrifugation and resuspended in cell lysis buffer containing 50 mM HEPES-KOH pH 8.0, 1 mM EDTA, 140 mM NaCl, 10% glycerol, 0.5% NP-40, 0.25% Triton X-100, and protease inhibitors. Cellular membranes were disrupted by Dounce homogenization. Samples were centrifuged for 10 minutes at 5000g at 4°C and resuspended in shearing buffer containing 10 mM Tris-HCl pH 8.0, 1 mM EDTA, 140 mM NaCl, 1.0% SDS, 0.1% Na deoxycholate, 1% Triton X-100, and protease inhibitors. Samples were then sonicated using the Fisher Scientific Model 120 Sonic Dismembrator (ThermoFisher Scientific) to obtain 200- to 500-bp fragments as confirmed by agarose gel electrophoresis. HepG2 cells were sonicated at 35% amplitude for 6 to 8 minutes with cycles of 30 seconds on and 30 seconds off, and UtLM cells were sonicated at 35% amplitude for 6 to 7.5 minutes with cycles of 30 seconds on and 30 seconds off. Sheared chromatin was precleared with protein A agarose/salmon sperm DNA (16-157; Millipore, Temecula, CA) and 20 µL sheared chromatin was removed as input DNA samples. Precleared samples were immunoprecipitated overnight with antibodies against GR or IgG. The DNA-protein complex was washed and precipitated with protein A magnetic beads (161-4013; BioRad), and cross-links were reversed. DNA was purified using the QiaQuick PCR Purification kit (28104; Qiagen). Immunoprecipitated DNA was quantified by qRT-PCR with custom-designed primer-probe sets as previously described (41).

Statistical analysis

Data represent the average of at least 4 biological replicates and are presented as the mean ± SEM. Statistical significance was determined by ANOVA with Tukey’s post hoc analysis using the version 7.01 GraphPad Prism software. Statistical significance is defined as P < 0.05 (*) or P < 0.01 (**).

Results

HepG2 and UtLM cells demonstrate robust GR transactivation in response to glucocorticoids

The objective of this study was to evaluate the impact of the selective progesterone receptor modulator ulipristal on glucocorticoid signaling in relevant human cell types. Therefore, we first established that the well-characterized HepG2 and UtLM cell lines were a suitable model for evaluating glucocorticoid activity (42, 43). The cellular actions of glucocorticoids are mediated by GR and, as previously demonstrated, HepG2 and UtLM cells express GR (Fig. 1A) (44, 45). Activation of GR by glucocorticoids results in phosphorylation at the serine 211 and 226 residues, which are associated with nuclear translocation and the transcriptional activity of GR (46–48). In both cell lines, acute treatment (1 hour) with the synthetic glucocorticoid dexamethasone resulted in phosphorylation at serine 211 and 226 of GR when compared with vehicle controls (Fig. 1B). GR transactivation in response to glucocorticoid treatment (synthetic and endogenous cortisol) resulted in the significant induction of 3 classic glucocorticoid-regulated genes: period circadian clock 1 (PER1), FK506 binding protein 5 (FKBP5), and glucocorticoid-induced leucine zipper (GILZ) (Fig. 1C) (49–51). Interestingly, the transcriptional response to glucocorticoids was more robust for all genes measured in UtLM cells compared with HepG2 cells. Collectively, these data suggest that the HepG2 and UtLM cell lines express GR, are responsive to glucocorticoid activation, and provide a relevant model to determine the effect of ulipristal on glucocorticoid activity in human liver and fibroid cell types.

Figure 1.

HepgG2 and UtLM cells are in vitro models of glucocorticoid responsiveness. (A) Representative western blot of GR and β-actin in untreated HepG2 and UtLM cells. (B) Representative western blot of GR phosphorylation at serine 211 (p211), serine 226 (p226), and β-actin in HepG2 and UtLM cells treated for 1.5 hours with Veh or 100 nM Dex. (C) Quantitative RT-PCR of mRNA isolated from HepG2 and UtLM cells treated for 6 hours with 1X PBS (black bars), 100 nM Cort (gray bars), or 100 nM (white bars). Transcript levels of PER1, FKBP5, and GILZ were normalized to the reference gene PPIB and set relative to Veh. Bar graphs represent the mean of at least 4 biological replicates ± SEM. **P < 0.01 compared with Veh as determined by ANOVA with Tukey’s post hoc analysis. Abbreviations: Cort, cortisol; dex, dexamethasone; UtLM, immortalized human uterine fibroid cell; Veh, vehicle.

Ulipristal blocks glucocorticoid-mediated gene transcription in HepG2 and UtLM cells

To determine whether ulipristal exhibits antiglucocorticoid activity in human liver and uterine fibroid cells, HepG2 and UtLM cells were pretreated with increasing concentrations of ulipristal (0, 100, and 1000 nM) for 30 minutes followed by treatment with vehicle (1X PBS) or 100 nM dexamethasone for 6 hours. The selective progesterone receptor modulator ulipristal alone did not exhibit agonist or antagonist activity on the endogenous expression of PER1, FKBP5, or GILZ mRNA in HepG2 (Fig. 2A) and UtLM cells (Fig. 2B). However, ulipristal pretreatment resulted in the dose-dependent repression of the response to dexamethasone in both cell lines. When treated at the same concentration (100 nM), ulipristal diminished the dexamethasone-mediated rise in transcript levels by 31% to 41% in HepG2 cells (Fig. 2A) and by 28% to 55% in UtLM cells (Fig. 2B). The transcriptional response to glucocorticoids was completely blocked by 1000 nM ulipristal in both cell lines. To evaluate whether the observed glucocorticoid antagonism by ulipristal compared with that of the known antiglucocorticoid RU-486 (mifepristone), cells were pretreated with RU-486 in the same manner as ulipristal, increasing concentrations (0, 100, and 1000 nM) for 30 minutes followed by treatment with vehicle (1X PBS) or 100 nM dexamethasone for 6 hours (32). In HepG2 cells, glucocorticoid antagonism by ulipristal was comparable to that of RU-486 (Fig. 2A). Of note, 100 nM RU-486 completely blocked the induction of PER1 mRNA by dexamethasone compared with the partial inhibition by ulipristal at the same concentration. The 1000 nM concentration of RU-486 partially suppressed the upregulation of GILZ by dexamethasone, whereas 1000 nM of ulipristal completely blocked the glucocorticoid effects. Inhibition of glucocorticoid signaling by ulipristal was also comparable to RU-486 in UtLM cells (Fig. 2B). Both 1000 nM RU-486 and 1000 nM ulipristal completely blocked the induction of PER1, FKBP5, and GILZ mRNA by dexamethasone in UtLM cells. These findings indicate that ulipristal is a potent inhibitor of glucocorticoid signaling in human liver and uterine fibroid cells, comparable to the antagonist activity of RU-486.

Figure 2.

Ulipristal antagonizes glucocorticoid signaling in HepgG2 and UtLM cells. Transcript levels of PER1, FKBP5, and GILZ mRNA measured by quantitative RT-PCR following the treatment of HepG2 (A) and UtLM (B) cells with 0, 100, or 1000 nM Uli, RU-486, or NOAC ± vehicle (1X PBS) or 100 nM Dex for 6 hours. Vehicle or Dex were added 30 minutes following Uli, RU-486, or NOAC treatment. Values were normalized to the reference gene PPIB and set relative to vehicle. Bar graphs represent the mean of at least four biological replicates ± SEM. Statistical significance when compared with vehicle (0 nM) was determined by ANOVA with Tukey’s post hoc analysis and is denoted by *P < 0.05 or **P < 0.01. Statistical significance between treatment groups is denoted as P < 0.05 (+) or P < 0.01 (++). Abbreviations: Dex, dexamethasone; NOAC, norethisterone acetate; Uli, ulipristal acetate; UtLM, immortalized human uterine fibroid cell.

Progestin-based treatments have been explored as long-term therapeutic options for various gynecological diseases (including uterine fibroids) (52–54). To evaluate whether the progestin NOAC could interfere with the cellular response to glucocorticoids, HepG2 and UtLM cells were treated with increasing concentrations (0, 100, and 1000 nM) of NOAC for 30 minutes followed by treatment with vehicle (1X PBS) or 100 nM dexamethasone for 6 hours. In HepG2 cells, pretreatment with 1000 nM NOAC resulted in a small but significant decrease in glucocorticoid-mediated PER1 transcript levels (Fig. 2A). NOAC did not alter the response to glucocorticoids for FKBP5 and GILZ (Fig. 2A). In UtLM cells, NOAC did not affect the transcriptional response to glucocorticoids. These results suggest that the antiglucocorticoid actions of medical therapeutics targeting gynecological diseases may be specific to those developed as antiprogestins.

To determine whether the timing of exposure (ie, ulipristal exposure before or following glucocorticoid activation) would affect the ability of ulipristal to antagonize glucocorticoid-mediated gene regulation, we treated HepG2 and UtLM cells concomitantly with 100 nM dexamethasone and 1 μM ulipristal or added 1 μM ulipristal 30 minutes following the treatment with 100 nM dexamethasone. Regardless of whether ulipristal treatment occurred at the same time or following Dex treatment in HepG2 (Fig. 3A) or UtLM (Fig. 3B) cells, the presence of ulipristal blunted the transcriptional response to glucocorticoids. In HepG2 cells, the expression of FKBP5 and GILZ mRNA was significantly greater than vehicle when cells were treated first with Dex and then ulipristal was added 30 minutes later, although this expression value was more than 60% less than cells treated with only dexamethasone (Fig. 3A). In UtLM cells, transcript levels of all 3 genes in cells treated with ulipristal and dexamethasone at the same time or with ulipristal following Dex were not statistically different from vehicle-treated cells (Fig. 3B). Therefore, ulipristal can inhibit glucocorticoid activity irrespective of the activation status of the glucocorticoid signaling pathway.

Figure 3.

Timing of ulipristal exposure does not alter glucocorticoid antagonist activity. Transcript levels of PER1, FKBP5, and GILZ mRNA measured by quantitative RT-PCR following the 6-hour treatment of (A) HepG2 and (B) UtLM cells with vehicle (1X PBS), 100 nM Dex, 1 μM Uli, 100 nM Dex + 1 μM Uli concomitantly (Co), or 100 nM Dex with 1 μM Uli added 30 minutes after Dex (After). Values were normalized to the reference gene PPIB and set relative to vehicle. Bar graphs represent the mean of at least 4 biological replicates ± SEM. Statistical significance when compared with vehicle was determined by ANOVA with Tukey’s post hoc analysis and is denoted **P < 0.01. Statistical significance between treatment groups is denoted as P < 0.01 (++). Abbreviations: Dex, dexamethasone; Uli, ulipristal acetate; UtLM, immortalized human uterine fibroid cell.

PR is not required for the antiglucocorticoid activity of ulipristal

Although glucocorticoids bind with lower affinity to the PR than GR, both Dex and ulipristal can bind PR (55, 56). Given that others have demonstrated PR-dependent activities of Dex in the mouse uterus, we tested whether PR was required for Dex-mediated gene regulation and the antiglucocorticoid actions of ulipristal in human cell types (57). Western blot analysis demonstrated detectable levels of PR in UtLM but not HepG2 cells (Fig. 4A). Moreover, PGR transcript levels were amplified at cycle number 28.9 ± 0.3 in UtLM cells but were undetectable by qRT-PCR following 40 cycles in HepG2 cells (data not shown). Thus, it is unlikely that the effects of Dex and ulipristal are mediated by PR in HepG2 cells. To determine the contribution of PR in UtLM cells, PR was depleted by siRNA-mediated gene silencing. Knockdown of PGR mRNA was evaluated by qRT-PCR in UtLM cells (Fig. 4B). Transfection of PGR siRNA significantly reduced expression of PGR compared with cells transfected with non-targeting control (NTC siRNA) but did not alter the expression of the GR gene (NR3C1). Depletion of cellular PR did not affect the ability of Dex to induce the expression of PER1, FKBP5, and GILZ mRNA (Fig. 4C). Moreover, PR knockdown did not relieve the inhibition of Dex-mediated gene expression by 1 μM ulipristal. Therefore, ulipristal inhibits glucocorticoid signaling in a PR-independent manner.

Figure 4.

Ulipristal inhibits glucocorticoid signaling in a PR-independent manner. (A) Representative western blot of PR and β-actin in untreated HepG2 and UtLM cells. (B) Transcript levels of PGR and NR3C1 mRNA measured by quantitative RT-PCR from UtLM cells transfected with NTC or PGR siRNA. Quantified mRNA values were normalized to the reference gene PPIB and set relative to the NTC siRNA sample. Statistical significance when compared with NTC siRNA was determined by a 2-tailed t-test and is denoted **P < 0.01. (C) Levels of PER1, FKBP5, and GILZ mRNA were measured by quantitative RT-PCR from UtLM cells transfected with NTC (black bars) or PGR (white bars) siRNA and treated for 6 hours with vehicle (1X PBS), 100 nM Dex, 1 μM Uli, or 100 nM Dex + 1 μM Uli. Cotreated cells were exposed to Uli for 30 minutes before the addition of Dex. Gene transcript values were normalized to the reference gene PPIB and set relative to the Veh NTC siRNA samples. Bar graphs represent the mean of 3 biological replicates ± SEM. Statistical significance when compared to Veh was determined by ANOVA with Tukey’s post hoc analysis and is denoted *P < 0.05 or **P < 0.01. Statistical significance between treatment groups is denoted as P < 0.01 (++). Abbreviations: Dex, dexamethasone; NTC, nontargeting control; siRNA, small interfering RNA; Uli, ulipristal acetate; UtLM, immortalized human uterine fibroid cell; Veh, vehicle.

Ulipristal blocks GR transactivation by inhibiting receptor phosphorylation, translocation, and DNA binding

To understand how ulipristal inhibits the transcriptional response to glucocorticoids, we initially evaluated the effect of ulipristal on the expression of GR protein levels. Treatment of HepG2 cells with 1 μM ulipristal for 24-hour treatment did not significantly alter GR levels (Fig. 5A). In contrast, GR expression was reduced by 40% in UtLM cells following 24-hour treatment with 1 μM ulipristal (Fig. 5B). Although the longer treatment length was associated with reduced GR protein in UtLM cells, transcript levels of the GR gene (NR3C1) were not reduced in HepG2 (Fig. 5C) or UtLM (Fig. 5D) cells following 6-hour treatment with 1 μM ulipristal. Moreover, GR protein levels were not significantly altered at the 6-hour timepoint in UtLM cells, when ulipristal treatment blocked glucocorticoid responsiveness (39). These data suggest that the mechanism by which ulipristal acutely blocks the transcriptional response to glucocorticoids is not attributed to altered GR expression.

Figure 5.

Ulipristal blocks glucocorticoid-mediated phosphorylation of GR. Representative western blot of GR and β-actin in (A) HepG2 and (B) UtLM cells treated for 24 hours with 1X PBS Veh or 1 μM Uli. Transcript levels of NR3C1 mRNA were measured by quantitative RT-PCR from HepG2 (C) and UtLM (D) cells treated for 6 hours with 0 nM, 100 nM, or 1000 nM Uli. Values were normalized to the reference gene PPIB and set relative to 0 nM. Bar graphs represent the mean of 4 or 5 biological replicates ± SEM. Statistical significance was evaluated by ANOVA. Representative western blot of GR phosphorylation at serine 211 (p211), GR, and β-actin in (E) HepG2 and (F) UtLM cells treated for 1.5 hours with vehicle (1X PBS), 100 nM Dex, 1 μM Uli, or Dex + Uli. Cotreated cells were exposed to Uli for 30 minutes before the addition of Dex. Representative western blot of GR phosphorylation at serine 226 (p226), GR, and β-actin in (G) HepG2 and (H) UtLM cells treated for 1.5 hours with vehicle (1X PBS), 100 nM Dex, 1 μM Uli, or Dex + Uli. Cotreated cells were exposed to Uli for 30 minutes before the addition of Dex. Abbreviations: Dex, dexamethasone; GR, glucocorticoid receptor; PBS, phosphate-buffered saline; Uli, ulipristal acetate; UtLM, immortalized human uterine fibroid cell; Veh, vehicle.

Site-specific phosphorylation of GR in response to glucocorticoid treatment regulates its transcriptional activity, demonstrated by various mutations to the phosphorylation sites of GR (58, 59). Therefore, we examined the phosphorylation status of GR at the glucocorticoid-responsive serine 211 and 226 in HepG2 and UtLM cells following treatment with vehicle (1X PBS), 100 nM dexamethasone, 1 μM ulipristal, or 100 nM dexamethasone and 1 μM ulipristal. Where cells were cotreated, ulipristal was added first for 30 minutes, followed by a 1-hour dexamethasone treatment. Phosphorylation increased rapidly at serine 211 in response to dexamethasone in HepG2 (Fig. 5E) and UtLM cells (Fig. 5F). However, pretreatment with ulipristal blocked glucocorticoid-mediated phosphorylation of serine 211. The phospho-S226 antibody also demonstrated considerable immunoreactivity toward GR in the Dex-treated HepG2 (Fig. 5G) and UtLM (Fig. 5H) cells. Comparatively, immunoreactivity was attenuated in cells treated with Dex and ulipristal. Thus, we found that ulipristal restricts glucocorticoid-dependent phosphorylation of GR at serine 211 and 226, which may contribute to the diminished response to Dex in the presence of ulipristal.

GR activity is also gated by its subcellular localization. It the absence of ligand, GR resides primarily in the cytoplasm, whereas ligand-binding promotes nuclear accumulation and transcriptional events (60). Ulipristal has been shown to inhibit the translocation of GR to the nucleus in COS-1 cells transfected with an expression vector for GR (38). To determine whether the presence of ulipristal restricts ligand-mediated nuclear translocation of endogenous GR in liver and uterine fibroid cells, we prepared nuclear extracts from HepG2 and UtLM cells treated for 1 hour with vehicle (1X PBS), 100 nM dexamethasone, and 1 μM ulipristal, or 100 nM dexamethasone and 1 μM ulipristal, where ulipristal was added 30 minutes before Dex. Glucocorticoid treatment promoted nuclear accumulation of GR in HepG2 (Fig. 6A) and UtLM cells (Fig. 6B). However, pretreatment with ulipristal prevented GR translocation to the nucleus. We visualized nuclear translocation by immunofluorescence using these treatment conditions in HepG2 (Fig. 6C) and UtLM (Fig. 6D) cells. Consistent with the cell fractionation studies, GR resided predominantly in the cytoplasm of vehicle-treated cells and was enriched in the nucleus of cells treated with Dex in both cell lines. GR was visualized to be primarily in the cytoplasmic compartment of cells treated with Dex and ulipristal.

Figure 6.

Ulipristal restricts nuclear translocation and DNA interactions of GR. Representative western blot of GR and β-actin from the nuclear fraction of (A) HepG2 and (B) UtLM cells treated for 1.5 hours with vehicle (1X PBS), 100 nM Dex, 1 μM Uli, or Dex + Uli. Cotreated cells were exposed to Uli for 30 minutes before the addition of Dex. Levels of β-actin were used as the loading control to normalize GR protein for quantification. Bar graphs represent the mean of at least 3 biological replicates ± SEM. Statistical significance when compared to Veh was determined by ANOVA with Tukey’s post hoc analysis and is denoted **P < 0.01. Representative image of GR (green) immunocytochemistry by confocal microscopy of (C) HepG2 and (D) UtLM cells treated for 1.5 hr with vehicle (1X PBS), 100 nM Dex, 1 μM Uli, or Dex + Uli. Cotreated cells were exposed to Uli for 30 minutes before the addition of Dex. Nuclei were visualized by DAPI staining (blue). Chromatin immunoprecipitation assays were performed with IgG (white bars) and GR (black bars) in (E) HepG2 and (F) UtLM cells treated for 1.5 hours with vehicle (1X PBS), 100 nM Dex, or Dex + 1 μM Uli. Cotreated cells were exposed to Uli for 30 minutes before the addition of Dex. Coimmunoprecipitated DNA was analyzed by quantitative RT-PCR using primers for a GR response element in the promoter of GILZ. Data are graphed as fold-change relative to Veh, and the bar graphs represent the mean of at least 4 biological replicates ± SEM. Statistical significance when compared with vehicle was determined by ANOVA with Tukey’s post hoc analysis and is denoted **P < 0.01. (G) Proposed model of the mechanisms by which Uli inhibits GR transactivation, which include preventing cytoplasmic phosphorylation, nuclear translocation, and DNA interactions. Abbreviations: Dex, dexamethasone; GR, glucocorticoid receptor; Uli, ulipristal acetate; Veh, vehicle.

We next evaluated whether the presence of ulipristal would prevent glucocorticoid-mediated DNA interactions. The induction of GILZ mRNA by glucocorticoids was inhibited by ulipristal (Fig. 2A & B). Therefore, we measured the association of GR at a previously reported glucocorticoid response element in the GILZ promoter by chromatin immunoprecipitation assays in HepG2 (Fig. 6E) and UtLM cells (Fig. 6F) (41). GR was strongly recruited to the GILZ promoter in response to Dex in both cell lines, but when cells were pretreated with ulipristal, recruitment of GR in response to Dex was lost. Therefore, ulipristal can inhibit glucocorticoid-mediated phosphorylation, nuclear translocation, and DNA interactions of GR (Fig. 6G). Collectively, these data are consistent with a model in which ulipristal functions as a competitive antagonist, preventing the glucocorticoid-stimulated transactivation of GR that regulates gene expression.

Ulipristal inhibits glucocorticoid-mediated transcription in primary fibroid tissue

To determine whether the observations in cell lines pertained to primary human tissue, uterine fibroid tissue was obtained from women undergoing elective gynecological surgery (Fig. 7A). Isolated fibroids were manually cut into tissue of equivalent size and cultured ex vivo on top of a transwell insert to allow media to surround the tissue (Fig. 7B). A portion of each fibroid sample was also evaluated by immunohistochemistry for the presence of GR. In agreement with previous staining of human uterine fibroid tissue, GR was expressed throughout the uterine fibroid samples (Fig. 7C) (45). Moreover, GR staining was primarily restricted to the nuclei, suggesting an activated receptor in the cells of the uterine fibroid tissue. The uterine fibroid tissue cultured ex vivo was treated for 6 hr with vehicle (1X PBS), 100 nM dexamethasone, and 1 μM ulipristal, or 100 nM dexamethasone and 1 μM ulipristal, where ulipristal was added 30 minutes before Dex. In primary uterine fibroid cells, glucocorticoid stimulation significantly increased the expression of PER1, FKBP5, and GILZ mRNA compared with vehicle, indicating that primary uterine fibroid tissue is glucocorticoid responsive (Fig. 7D). However, in fibroid tissue pretreated with ulipristal, the induction of these genes by glucocorticoids was abolished. These data suggest that ulipristal strongly inhibits glucocorticoid responsiveness in primary human tissue.

Figure 7.

Ulipristal antagonizes glucocorticoid signaling in primary human uterine fibroid explants. (A) Subject tissue was isolated from 4 women undergoing elective gynecological surgery for benign conditions. Subject demographics are listed. (B) Representative image of uterine fibroid tissue dissected into approximately equal size fragments and cultured ex vivo using transwell inserts. (C) A portion of each tissue was fixed for histology immediately following dissection. Representative images of GR immunohistochemistry counterstained with hematoxylin in subject uterine fibroid tissue. Negative control image demonstrates staining in the absence of primary antibody. (D) The remaining tissue was dissected to approximately 30-mg fragments and placed in culture medium for ex vivo treatments. Transcript levels of PER1, FKBP5, and GILZ mRNA were measured by quantitative RT-PCR in human uterine fibroid explants treated ex vivo for 6 hours with 1X PBS Veh, 100 nM Dex, 1 µM Uli, or Dex + Uli, where Uli was added 30 minutes before Dex. Quantified mRNA values were normalized to the reference gene PPIB and set relative to Veh. Bar graphs represent the mean of 4 biological replicates ± SEM. Statistical significance when compared to vehicle was determined by ANOVA with Tukey’s post hoc analysis and is denoted **P < 0.01. Dex, dexamethasone; GR, glucocorticoid receptor; Uli, ulipristal acetate; Veh, vehicle.

Ulipristal demonstrates GR antagonist and agonist properties in vivo

To define the physiological significance of ulipristal on glucocorticoid signaling in vivo, we exposed female C57Bl/6 mice to 2 mg/kg ulipristal for 4 hours. The dose chosen for exposure was based on clinical trial data in which women received ulipristal for the treatment of uterine fibroids and scaled appropriately accounting for the differences in body surface area in mice compared with humans (17, 18, 40, 61). Tissues were isolated for gene expression analysis based on demonstrated GR actions and included the lung, spleen, liver, uterus, hippocampus, and pituitary (62). Acute exposure to ulipristal resulted in tissue-specific effects on basal glucocorticoid signaling, including antagonist, agonist, and no effects. In the lung, ulipristal significantly reduced transcript levels of Per1 and Fkbp5 but not Gilz (Fig. 8A). Consistent with the findings in Hepg2 and UtLM cells, transcript levels of Per1, Fkbp5, and Gilz were significantly decreased by ulipristal in the liver and the uterus (Fig. 8A). However, no effect of ulipristal exposure was evident on the basal expression of the glucocorticoid-responsive genes Per1, Fkbp5, and Gilz in the spleen (Fig. 8B). Interestingly, exposure to ulipristal induced the expression of Fkbp5 and Gilz mRNA in the hippocampus and pituitary (Fig. 8C), suggesting that ulipristal may act as an agonist of glucocorticoid signaling in certain tissues.

Figure 8.

Ulipristal alters glucocorticoid signaling in a tissue-specific manner in mice. Adult female C57Bl/6 mice were injected with corn oil Veh or 2 mg/kg Uli for 4 hours. mRNA levels for Per1, Fkbp5, and Gilz were determined by quantitative RT-PCR from isolated tissues. Quantified mRNA values were normalized to the reference gene Ppib and set relative to Veh. Bar graphs represent the mean of 3 to 7 mice per group ± SEM. Statistical significance when compared with Veh was determined by 2-tailed t-test and is denoted *P < 0.05 or **P < 0.01. Patterns in gene regulation were grouped by effect. (A) Uli inhibited the basal expression of glucocorticoid responsive genes in the lung, liver, and uterus. (B) Exposure to Uli did not alter the expression of Per1, Fkbp5, and Gilz in the spleen. (C) Uli induced the expression of glucocorticoid responsive genes in the hypothalamus and pituitary. Abbreviations: Uli, ulipristal acetate; Veh, vehicle.

Discussion

The clinical pause in the long-term application of ulipristal provides an opportunity for further investigation into potential unappreciated, off-target effects. Unanticipated adverse events may reflect an incomplete understanding of the mechanisms by which this and potentially other PR modulator act. In fact, our data demonstrate that ulipristal exhibits potent GR antagonist activity in human liver and uterine fibroid cells, completely blocking glucocorticoid-induced gene transcription. In uterine fibroid and liver cells, the suppression of glucocorticoid signaling with ulipristal was comparable to that of the known GR antagonist RU-486. Moreover, we found that this effect was not dependent on the presence of PR, suggesting a direct inhibition of glucocorticoid signaling through GR. Glucocorticoid-mediated responses reflect the ligand-dependent transactivation of GR, which is characterized by receptor phosphorylation, nuclear translocation, and DNA binding (62). We determined that the mechanism by which ulipristal antagonizes GR is by inhibiting these critical steps of GR transactivation. These data are consistent with findings by Wagner et al., who described RTI 3021-012 (ulipristal) as a competitive antagonist for GR, preventing nuclear translocation and DNA binding in transfection assays (38). We have determined that ulipristal blocks the activity of endogenous GR in human cell types relevant to its potential clinical application as a long-term therapeutic agent.

Glucocorticoids were originally named for their role in the regulation of hepatic gluconeogenesis and, accordingly, the liver is a major target of glucocorticoid action (63). The physiological role for glucocorticoid signaling in the liver has been demonstrated with various liver-specific GR knockout mouse models (62). Hepatocyte-specific inactivation of GR early in embryonic development critically impairs postnatal growth and survival (64). Approximately 50% of pups born without hepatic GR die within 48 hours of birth; in surviving adults, the absence of hepatic GR results in significant hypoglycemia in response to metabolic challenge (65). The association between disrupted hepatic glucocorticoid signaling and metabolic defects may reflect altered regulation of circadian regulatory genes, including the GR targets Per1 and Per2. The molecular circadian clock mediates many aspects of liver physiology, including glucose regulation, lipid synthesis/metabolism, and bile acid production. The importance of direct regulation by GR was demonstrated in a mouse model containing a selective deletion of a glucocorticoid response element within Per2 (66). Simply deleting the ability of Per2 to be regulated by glucocorticoids resulted in altered glucose homeostasis in these mice, demonstrating that disruptions to GR-mediated transcription in the liver can result in pathophysiology.

Characterization of the liver-specific GR knockout mice also revealed a protective role for GR in the development of liver disease (67). Correspondingly, abnormal liver enzymes, including elevated aspartate aminotransferase and alanine aminotransferase, were reported in patients with subclinical Addison disease (68). Importantly, liver enzyme levels in these patients normalized following glucocorticoid replacement therapy. Altered liver functions have also been reported following treatment with GR antagonists. Liver toxicity was described for 2 women with Cushing disease taking daily RU-486, which resolved following the discontinuation of treatment (35, 69). Although RU-486 is categorized by the National Institutes of Health as unlikely to cause clinically apparent liver injury, this classification was based on the short-term use of RU-486 to induce abortion and may not reflect the impact of protracted exposure. Rats treated with long-term RU-486 and metapristone, the major metabolite of RU-486, exhibited hepatotoxicity and elevated but reversible levels of alanine aminotransferase and aspartate aminotransferase, specifically in females (36). Therefore, the association between repressed glucocorticoid signaling and liver injury is important and well-supported.

The activity of an agonist or antagonist is dependent on its affinity for the target receptor, but ligand binding is only partially responsible for dictating the physiological effects. For example, the selective estrogen receptor modulator tamoxifen acts as a receptor antagonist in the breast but has partial to full agonist activity in the uterus and bone, revealing that the cellular context is a key determinant of the response to ligand binding (70–72). Contextual plasticity in response to ligands has also been described for PR and GR (73, 74). Studies with mouse mammary and pituitary corticotroph cells demonstrated that the genes regulated by glucocorticoids and genomic regions bound by GR are considerably different across cell types (75, 76). In fact, only 11.4% of GR-bound genomic regions were shared between these 2 cell types. In addition to differences in chromatin interactions, the expression of transcriptional cofactors varies by cell, which can shift the balance between upregulation and downregulation. Using the Human Protein Atlas, several common coactivators and corepressors were found to be differentially expressed in the liver compared with hypothalamus and pituitary, including the nuclear receptor coactivator-2, -4, -5, -6, and -7 and nuclear corepressor 1 (77). The unique properties of GR also allow for tissue-selective actions, including the presence of transcriptional splice variants, translational isoforms, and posttranslational modifications (74, 78, 79). The cellular distribution of these GR subtypes differs by tissue and disease states.

Our current work indicates that ulipristal has mixed agonist and antagonist properties in vivo depending on the tissue evaluated. We found that ulipristal antagonizes the basal expression of classic glucocorticoid-responsive genes Per1, Fkbp5, and Gilz in the mouse lung, liver, and uterus. Unexpectedly, we discovered that ulipristal induces the expression of Fkbp5 and Gilz in the hypothalamus and pituitary, suggesting GR agonist activity in the brain. Glucocorticoid exposure produces many diverse effects in the hypothalamus and pituitary, including increased ghrelin and neuropeptide Y expression and decreased release of prolactin, thyroid-secreting hormone, corticotropin-releasing factor, and gonadotropin-releasing factor, resulting in an overall suppression of hypothalamic and pituitary activity (80–83). Whether ulipristal possess tissue-specific GR agonist activity in the human brain is unknown but is an important consideration because of the potential to disrupt the hypothalamic-pituitary-adrenal axis.

The family of steroid hormone receptors, to which PR belongs, evolved from a common ancestral protein and maintains a well-conserved ligand binding domain (33). Within this family, PR, GR, the androgen receptor (AR), and the mineralocorticoid receptor (MR) belong to a subclass of paralogous receptors able to bind diverse non-aromatized steroids. Because of the evolutionary relatedness of PR, GR, MR, and AR, it is possible that ulipristal can modulate the activity of other members of this receptor subclass. Preclinical studies reported that ulipristal can bind AR, albeit with much lower affinity than the endogenous ligand, and demonstrates modest antiandrogenic effects in the rat prostate (37, 84). In addition to playing an important role in the development and functions of the reproductive tract, AR signaling regulates metabolism, immune cell development, bone homeostasis, and cardiovascular function (85, 86). Global and liver-specific AR knockout mouse models have demonstrated that AR plays a protective role in regulating liver physiology (87). In the absence of hepatic AR, male mice develop liver steatosis, which is exacerbated by high-fat diet. Similar to RU-486, acute liver injury has also been reported for patients taking AR antagonists (88–90).

Interestingly, AR expression is upregulated in uterine fibroid tissue compared with normal myometrium, although the role of androgen signaling in uterine fibroids is unclear (91). The presence of the non-sex steroid hormone receptors may represent unexplored signaling pathways contributing to the pathogenesis of uterine fibroids. We have previously demonstrated that glucocorticoids can preferentially signal in uterine fibroid cells compared to normal myometrial cells in vitro, where the transcript levels of certain genes were uniquely regulated in uterine fibroid cells (45). Although the genes evaluated in this study are not currently associated with uterine fibroid pathophysiology, the implication of active GR signaling in uterine fibroid cells is important. In the presence of relatively equivalent levels of GR, we found that glucocorticoids induced a more robust transcriptional response in the uterine fibroid cells compared to HepG2 cells. This suggests that the unique cellular context of uterine fibroid cells supports a highly responsive GR. The posttranslational modifications to GR act as allosteric regulatory factors conferring context-specific activity (79). In this study, we determined that glucocorticoid-mediated phosphorylation of GR at serine 226 was greater in the uterine fibroid cells compared with HepG2 cells. Further work is needed to determine whether there are also differences in the allosteric regulators of GR when comparing normal myometrium and uterine fibroids, which may contribute to the pathogenesis of this disease. Interestingly, we found that long-term ulipristal treatment downregulated the expression of GR in uterine fibroid but not HepG2 cells. It is unclear whether ulipristal-dependent changes in GR expression contribute to the mechanism by which this therapeutic reduces uterine fibroid volume and symptoms, as the actions of GR in uterine fibroids have not been defined. Moreover, the molecular mechanisms that regulate the development and growth of uterine fibroids are not well understood. Therefore, it will be important for future studies to examine the role of GR in promoting uterine fibroid growth.

The pursuit for compounds that inhibit PR without altering the biological actions of glucocorticoids is paramount to the treatment of gynecological diseases but continues to challenge the field. This study provides a format for evaluating the antiglucocorticoid effects of the SPRMs. Using human cell lines and primary human tissue, we provide evidence that the selective PR modulator ulipristal can block the cellular response to glucocorticoids in uterine fibroid and liver cells by acting as a competitive antagonist for GR. Much like the selective estrogen receptor modulators, data from our in vivo studies demonstrate that ulipristal exhibits tissue-selective antagonist and agonist effects on GR. Therefore, our findings indicate that the limited GR modulating activity of ulipristal determined by receptor modeling and reporter assays may not reflect the effect on the endogenous receptor and have been underestimated (37, 92). These unexpected results support a broader characterization of the selective receptor modulators before long-term use in humans. Fully understanding the molecular mechanisms by which the selective receptor modulators function will advance the development of effective, long-term treatment options with limited, unanticipated side effects.

Glossary

Abbreviations

AR

androgen receptor

Dex

dexamethasone

FBS

fetal bovine serum

GR

glucocorticoid receptor

MEM

minimum essential medium

MR

mineralocorticoid receptor

NGS

normal goat serum

NOAC

norethisterone acetate

NTC

nontargeting control

pGR

antiphosphorylated glucocorticoid receptor

PR

progesterone receptor

qRT-PCR

quantitative RT-PCR

siRNA

small interfering RNA

SPRM

selective progesterone receptor modulator

TLC

thin-layer chromatography

Uli

ulipristal acetate

UtLM

immortalized human uterine fibroid cell

Additional Information

Disclosure Summary: Ayman Al-Hendy is a consultant for Bayer, Myovant Sciences, and AbbVie. All other authors have nothing to disclose.

Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.