Activation of β-Catenin Signaling and its Crosstalk With Estrogen and Histone Deacetylases in Human Uterine Fibroids

Abstract

Context

Uterine fibroids (UF) are the most common benign tumor of the myometrium (MM) in women of reproductive age. However, the mechanism underlying the pathogenesis of UF is largely unknown.

Objective

To explore the link between nuclear β-catenin and UF phenotype and β-catenin crosstalk with estrogen and histone deacetylases (HDACs).

Design

Protein/RNA levels of β-catenin (CTNNB1 gene), its responsive markers cyclin D1 and c-Myc, androgen receptor (AR), p27, and class-I HDACs were measured in matched UF/MM tissues or cell populations. The effects of chemical inhibition/activation and genetic knockdown of CTNNB1 on UF phenotype were measured. The anti-UF effect of 2 HDAC inhibitors was evaluated.

Main Outcome Measure

β-catenin nuclear translocation in response to β-catenin inhibition/activation, estrogen, and HDAC inhibitors in UF cells.

Results

UF tissues/cells showed significantly higher expression of nuclear β-catenin, cyclin D1, c-Myc, and HDACs 1, 2, 3, and 8 than MM. Estradiol induced β-catenin nuclear translocation and consequently its responsive genes in both MM and UF cells, while an estrogen receptor antagonist reversed this induction effect. Treatment with β-catenin or HDAC inhibitors led to dose-dependent growth inhibition, while Wnt3a treatment increased proliferation compared with control. Chemical inhibition of β-catenin decreased cyclin D1 and c-Myc expression levels, while β-catenin activation increased expression of the same markers. Genetic knockdown of CTNNB1 resulted in a marked decrease in β-catenin, cyclin D1, c-Myc, and AR expression. Treatment of UF cells with HDAC inhibitors decreased nuclear β-catenin, cyclin D1, and c-Myc expression. Moreover, HDAC inhibitors induced apoptosis of UF cells and cell cycle arrest.

Conclusion

β-catenin nuclear translocation contributes to UF phenotype, and β-catenin signaling is modulated by estradiol and HDAC activity.

Uterine fibroids (UFs), also known as myomas or leiomyomas, are the most prevalent benign smooth muscle tumors of the uterus. The lifetime risk among women for the development of UFs by the onset of menopause is as high as 70% to 80% in the United States (1). Symptoms occur in 25% to 50% of affected women, causing a significant clinical burden, including abnormal uterine bleeding, pelvic pain, pregnancy loss, and infertility. UF pose an enormous economic burden worldwide (2). However, the mechanism underlying the pathogenesis of UFs remains unclear. An increasing body of evidence supports the hypothesis that these monoclonal tumors originate from aberrant stem cells in the myometrium (MM) and involve multiple factors, including the sex steroid hormones estrogen and progesterone, various growth factors, and inherent/genetic traits (3). Recent studies have demonstrated that somatic mutations in the mediator complex subunit 12 (MED12) gene exon 2 occur at high frequency in UFs (80%) in women from diverse racial and ethnic origins; these mutations are a major genetic abnormality contributing to UF pathogenesis and have been linked with activation of the Wnt/β-catenin signaling pathway (4, 5).

The Wnt/β-catenin signaling pathway is well known to be involved in a variety of biological events, such as development, tissue renewal, and cell proliferation and differentiation, as well as in several types of carcinogenesis (6, 7). Cellular β-catenin resides in 3 different areas: membranous β-catenin (which maintains cell adhesion), cytoplasmic, and nuclear. Cytoplasmic β-catenin is unstable and rapidly targeted for degradation by a large multi-protein complex composed of Axin1, adenomatous polyposis coli, glycogen synthase kinase3β (GSK3β), casein kinase-Iα (CK-Iα) and protein phosphatase 2A. Activation of this pathway occurs via the interaction of a Wnt ligand with the Frizzle/LRP co-receptor complex, which prevents β-catenin degradation and increases β-catenin translocation into the nucleus, increasing the activity of the TCF/LEF transcription complex and triggering the gene expression (8). In vitro data suggest that targeting the Wnt/β-catenin pathway may be useful in poor-prognosis neoplasia, such as triple-negative breast cancer (9). Our group and others have previously demonstrated that inhibiting Wnt/β-catenin pathway using vitamin D3 is associated with decreased UF cell proliferation (10, 11).

Nuclear accumulation of β-catenin within the cells has been shown in many types of cancer types and is related to increased tumorigenicity (12–14). Moreover, higher cytoplasmic expression followed by nuclear localization of β-catenin is characteristic of stem-like cell populations in chemotherapy-resistant cancers and capable of initiating new tumors (15).

Involvement of Wnt signaling in UF formation has been increasingly described in recent years (4, 10, 16, 17). However, there is controversy regarding the expression of β-catenin in UFs compared with normal MM, with studies suggesting either upregulation of β-catenin expression in UFs (18) or no difference of expression between UFs and MM (19). Notably, there is limited knowledge about the localization of β-catenin in UF cells and its contribution to disease pathophysiology. β-catenin nuclear translocation activates the transcription of its pro-growth responsive genes such as CCND1 and c-Myc and increases cell proliferation, cell invasion, and DNA repair (14, 20, 21).

As reported in the literature, both β-catenin and estrogen signaling represent important pathways in UF growth promotion (1, 16, 22). An increasing number of observations suggest a potential convergence between these pathways in several tumors, such as colon and breast cancers, as well as in endometriosis and neurodegenerative disease (23–28). Several studies reported a hypothesized role of estrogen in activation of β-catenin in UF-like lesions (5, 29) and uterine epithelium (24). Moreover, estrogen induced paracrine activation in UF (16). However, the functional interaction to promote UF growth has not been reported.

Tumorigenesis is not explained solely via genetic changes, but also involves epigenetic processes (30, 31). Acetylation of histones, which can play a key role in epigenetic regulation of gene expression, is controlled through a balance between histone deacetylases (HDACs) and histone acetyltransferases (HATs) (31). HDAC inhibitors have been shown to induce cancer cell–cycle arrest and cell death (30). Interestingly, the acetylation status of several nonhistone proteins, including β-catenin, estrogen receptor (ER) and c-Myc, can modify many cellular functions, such as mRNA splicing, transport, and integrity, as well as translation, activity, localization, stability, and protein interactions (32, 33). Co-suppression of Wnt/β-catenin, HDAC, and ER has effectively repressed both bulk and cancer stem cell subpopulations in hormone-dependent breast cancer (34). HDAC activity was shown to be increased in UF primary cells, compared with MM cells, after treatment with estrogen (35), suggesting crosstalk between estrogen signaling and HDAC activity and that higher activity of HDAC could be involved in transcriptional repression of tumor suppressor genes such as p21 and p53 (35, 36), contributing to the maintenance and growth of UFs.

The mechanism underlying the regulation of β-catenin through estrogen and HDACs in UFs is largely unknown. Our hypothesis is that increased nuclear translocation of β-catenin contributes to the UF phenotype by activating its regulated genes. In addition, estrogen and HDACs play a critical role in regulation of β-catenin signaling.

Materials & Methods

Human tissue sample collection and primary cell isolation

Freshly collected human UF and adjacent MM samples were obtained via the Augusta University Biorepository, under approved an IRB protocol (IRB No. 644354–6), from consented women of reproductive age (22–55 years) who were undergoing hysterectomy or myomectomy for symptomatic UFs. These patients had not taken any hormonal supplements for 3 months prior to the day of surgery (ie, the day of sample collection). An 8-cm³ UF tissue sample was collected from each patient. Myometrial tissue samples were collected from at least 2 cm distance from adjacent UF to exclude any mechanical or hormonal effects of UFs on adjacent MM tissue. For preparation of primary cell populations, collected samples were washed with calcium- and magnesium-containing Hanks’ balanced salt solution (HBSS) to remove blood. After the tissues were chopped into small pieces, they were then digested overnight at 37°C by shaking in an enzyme buffer of calcium- and magnesium-free HBSS containing 1% antibiotic-antimycotic, 2.5% N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid [HEPES], 0.66 mg/mL collagenase type IV (Worthington, New Jersey), and 4.76 µg/mL DNase I (Sigma-Aldrich, St. Louis, Missouri). The suspension was then filtered through 100-µm sterile nylon mesh cell strainer to remove undigested tissues and refiltered through a 70-µm cell strainer (BD-Falcon) to obtain a single-cell suspension. Cells were plated out and incubated at 37°C, enabling the cells to attach to the sterile tissue culture treated plate containing smooth muscle basal medium (SmBM) culture media.

Regents and antibodies

17 β-estradiol and β-catenin inhibitors (cordycepin, XAV939) were purchased from Sigma Biochemicals (St. Louis, Missouri), and ICG-001 was from Selleckchem (Houston, Texas). Beta-catenin activator Wnt3a was purchased from R&D Systems (Minneapolis, Minnesota). Estrogen receptor antagonist, ICI182780, propidium iodide and HDAC inhibitors (HDAC inhibitor VIII and apicidin) were obtained from Millipore (Burlington, Massachusetts) and SmBM was purchased from Lonza (Walkersville, Maryland). Dimethyl sulfoxide (DMSO), phosphate-buffered saline (PBS), and DMED/F-12 were purchased from Gibco/Thermo Fisher Scientific (Waltham, Massachusetts). The apoptosis detection kit was from Biolegend (San Diego, California). Antibodies used in this study are listed in Reference) (37).

Cell lines and cultures

The immortalized human UF cell line (HuLM) and immortalized human uterine smooth muscle cells (UTSM) were a generous gift from Dr. Darlene Dixon (National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina). These cells were cultured and maintained in SmBM culture medium with 5% fetal bovine serum (FBS), 0.1% insulin, 0.2% recombinant human fibroblast growth factor (hFGF-B), 0.1% gentamicin sulfate and amphotericin B mixture and 0.1% human epidermal growth factor (hEGF) at 37°C in a humidified atmosphere of 5% CO₂/95% air. HuLM cells were passaged 2 times and once they were 70% confluent, the media was replaced by phenol red-free, serum-free DMEM/F-12 for different treatments and durations as described in the figure legends. Cell lines or isolated primary cells were trypsinized and washed in PBS, and centrifuged at 1500 revolutions per minute for 5 minutes. The supernatant was removed, and pellets were snap-frozen in liquid nitrogen and stored at −80°C until further molecular analysis.

Protein extraction and Western blot analysis

Pellets were lysed in lysis buffer with protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific, Waltham, Massachusetts), protein was quantified using Bradford method (Bio-Rad protein Assay kit, Hercules, California), 30 µg of protein lysates were resolved in Gradient (4%–20%) MiniPROTEAN TGX Precast Protein Gels (Bio-Rad, Hercules, California), transferred onto a polyvinylidene fluoride (PVDF) membrane (Bio-Rad), blocked for 1 hour at room temperature in either 5% weight/volume nonfat dry milk or 5% bovine serum albumin (BSA) in 0.05% Tween-supplemented PBS (0.05% PBST) per antibody specification and incubated with specific primary antibodies as listed in Reference (37) overnight at 4°C, followed by 90 minutes incubation with appropriate HRP-conjugated secondary antibodies (Cell Signaling Technology, Danvers, Massachusetts). Antigen-antibody complex was detected with the Pierce Enhanced Chemiluminescence (ECL) detection kit (Thermo Fisher Scientific) or the Trident femto Western HRP substrate (GeneTex, Irvine, California). Specific protein bands were visualized using the ChemiDoc imaging system (Bio-Rad, Hercules, CA). The intensity of each protein band was quantified by Bio-Rad Image Lab software and normalized against corresponding β-actin; normalized values were used to create data graphs.

Nuclear and cytoplasmic fractionation

Nuclear-cytoplasmic fractionation was conducted using the NE-PER Nuclear and Cytoplasmic Extraction Reagents kit (Thermo Fisher Scientific) according to the manufacturer’s instructions.

Isolation of cellular RNA and cDNA synthesis

Total cellular RNA was isolated from frozen pellets using the PureLink RNA Mini Kit (Ambion). The concentration of total RNA was determined using NanoDrop (Thermo Scientific). One microgram of total RNA from each sample was reverse-transcribed in 20 µl reaction volume to complementary DNA (cDNA) using Ecodry premix double primed (Clontech lab). The reaction mixture was incubated for 1 hour at 42°C and stopped by incubation at 70°C for 10 minutes.

Quantitative real-time polymerase chain reaction

Quantitative real-time polymerase chain reaction (qRT-PCR) was performed to determine the mRNA expression of several genes listed with their primer sequences in Reference (37). Primers were purchased from Integrated DNA Technologies (Coralville, Iowa). An equal amount of cDNA from each sample was added to the Mastermix containing appropriate primer sets and SYBR green supermix (Bio-Rad) in a 20 µl reaction volume. All samples were analyzed in triplicates. Real-time PCR analyses were performed using a Bio-Rad CFX96. Cycling conditions including denaturation at 95°C for 2 minutes followed by 40 cycles of 95°C for 5 seconds and 60°C for 30 seconds, followed by 65°C for 5 seconds. Synthesis of a DNA product of the expected size was confirmed by melting curve analysis. 18S ribosomal RNA values (internal control) were used to normalize the expression data and normalized values were used to create data graphs. Negative control was performed by running the reaction without cDNA template.

Immunofluorescence and laser confocal microscopy

A total of 8 ×10⁴ HuLM cells were cultured over sterile glass cover slips in 6-well plates; when the cells reached 60% confluence, they were serum-starved overnight and treated with ICG-00, apicidin, or HDAC inhibitor VIII (HDACi VIII) for 24 hours. DMSO vehicle was added to the control cells. Cells were fixed in a 4% formaldehyde solution at room temperature for 15 minutes. After washing in PBS 3 times, cells were permeabilized for 15 minutes using 0.1% Triton X-100/PBS, and then nonspecific binding was inhibited by blocking for 1 hour in blocking/incubation solution containing 1% bovine serum albumin (BSA) in 0.1% Triton X-100/PBS. Incubation with primary antibodies (37) was performed for 2 hours followed by incubation with anti-rabbit Alexa Fluor 555-conjugated secondary antibody (Thermo Fisher) (1:1000 each) for 1 hour at room temperature. Cells were washed for 15 minutes (3 washes of 5 minutes each) with the above-described incubation solution, air-dried, and mounted onto microscopic slides with a drop of Fluorshield (Sigma) containing 40,6 diamidino-2-phenylindole (DAPI) for nuclear staining. Negative controls were stained with secondary antibody only. Fluorescent signals were visualized using Zeiss 780 upright laser confocal fluorescent microscopy and ZEN Black 2012 confocal software. Images were captured at with regular 20× or 40× magnification using 40× Plan-Apo (oil)/1.4 numerical aperture lens and exported using Zen Blue 2012 software.

Cell proliferation assay

Cell proliferation was evaluated by dimethylthiazolyl diphenyltetrazolium bromide (MTT) assay. Briefly, 2000 HuLM cells were seeded in 96 well plates and treated with the different treatments for various time points as described in the figure legends. DMSO vehicle was added to the control cells. Each well received 100 µl of 0.5 mg/mL MTT in PBS solution and the plates were incubated at 37°C for 4 hours. They were then aspirated, 150 µl of DMSO was added to each well, and the plates were gently agitated on a shaker for 15 minutes while protected from light. Absorbance was measured in a Synergy HT multi-detection microplate reader (BioTek, Broadview, Illinois) at 570 nm.

Immunohistochemistry

Collected UF and adjacent MM samples were fixed in 10% buffered formalin for 15 to 20 hours, embedded with paraffin and subjected to immunohistochemistry (IHC). Paraffin-embedded tissue sections were deparaffinized and rehydrated by being passed through xylene and graded ethanol solutions as previously described. Nine matched pairs were stained for β-catenin, 3 pairs were stained for cyclin D1 and c-Myc (37) and subsequently stained with peroxidase-conjugated secondary antibody. For visual evaluation, each sample was scored for staining intensity, and the positive cells were semi-quantified. To determine the protein immunostaining score, we used a design proposed elsewhere (38). In brief, the semi-quantitative score was established by considering the percentage of labeled cells (0, negative; 1, < 10% of the cells; 2, 10%-50% of the cells; 3, 50%-75% of the cells; 4, > 75% of the cells) and the intensity of the immunostaining (0, no staining; 1, weak; 2, mild; 3, strong staining). The multiplication of both scores resulted in a final quotient ranging from 0 to 12. An independent pathologist performed a double-blind analysis.

Generation of β-catenin knockdown cells

Transient knockdown of β-catenin (CTNNB1) expression was achieved using human short hairpin (shRNA) lentiviral particles (OriGene). Briefly, at day 1, HuLM cells were cultured in a 12-well plate. Culture media was changed the following day with fresh media containing the lentiviral particles (MOI = 10) and polybrene (8 µg/mL) to enhance the transduction efficiency for 18 to 20 hours, then the medium was changed again and cells were left to grow until day 5. Protein lysate was isolated for measurement of protein levels in HuLM with either CTNNB1 human shRNA lentiviral particles or scramble control.

Apoptosis assay

HuLM cells were treated with apicidin or HDACi VIII for 24 hours at the indicated concentrations. Cells were washed with PBS and collected by trypsinization and then resuspended in Annexin V binding buffer. Annexin V-APC (conjugated with allophycocyanin) and 7-aminoactinomycin D (7-AAD) cell viability stains were added followed by incubation of the cells for 15 minutes at room temperature in the dark. Cells were washed with binding buffer. The population of apoptotic cells was analyzed with a BD FACS LSR Fortassa flow cytometer. Staurosporine was used as the positive control.

Cell cycle analysis

HuLM cells were treated with apicidin or HDACi VIII for 24 hours using the indicated concentrations. Cells were then harvested, washed with PBS, and fixed with 70% ethanol overnight at −20 °C. The cells were washed in PBS and incubated with DNA staining solution [10 mg/mL DNase free RNase (Thermo Fisher) and 50 µg/mL propidium iodide at room temperature for 30 minutes in the dark. Cells were then stored at 4°C and analyzed the following day by flow cytometry. The results are presented as the percentage of cells in each cell cycle phase.

Statistical Analysis

For quantitative RT-PCR data, gene expression results were presented as fold change ± standard error of mean (SEM) and statistical analysis were calculated using Bio-Rad CFX Manager 3.1 software; experiments were performed in triplicate and repeated twice. For Western blot analysis, results were presented as either mean ± standard deviation (SD) or SEM and analyzed using 2-tailed unpaired Student t-test, experiments were performed in triplicate and repeated twice. Differences were considered statistically significant if P < 0.05. Graph Pad 7.0 (La Jolla, California) was used for generating the graphs.

Results

Nuclear translocation of β-catenin is responsible for activation of its signaling pathway in human uterine fibroid tissue

First, we explored the protein expression of β-catenin and its responsive markers cyclin D1 and c-Myc, as well as androgen receptor (AR) and tumor suppressor p27 in 7 matched human UF and MM tissues (Fig. 1A). Total expression of β-catenin was overall significantly higher in UF compared with MM (P = 0.0391) (Fig. 1B); however, we noticed a patient-to-patient variability, as the expression level was higher in UF in 5 patients while unchanged or lower in 1 patient each. Interestingly, the expression of β-catenin–responsive pro-growth markers cyclin D1 and c-Myc was significantly higher in UF compared with MM (P = 0.0156 and 0.0078 respectively), even in the patients that showed either lower β-catenin expression in UF (patient #3 for cyclin D1) or no changes of β-catenin between UF and MM (patient #4 for c-Myc). β-catenin has been reported to form a complex with AR (39), and evolving evidence indicates a role for androgens and their receptors in UF pathogenesis (40, 41). In this study, AR expression was significantly higher in UF (P = 0.03) while p27 was lower in UF in 6 patients, reflecting higher proliferation in UF compared with MM tissues.

Figure 1.

Expression of β-catenin and its responsive genes in human uterine fibroid (UF) and myometrium (MM) tissues. ( A) Western blot analysis of protein lysates from matched UF and MM tissues (n = 7) for β-catenin, proteins encoded by β-catenin-regulated genes (cyclin D1, c-Myc, and androgen receptor [AR]), and the protein product (p27) of the tumor suppressor gene p27. (B) Graphical representation for densitometry quantification of the results obtained by Western blot; the intensity of each protein signal was quantified and normalized to the corresponding β-actin and presented in the graph as mean ± SEM. Experiments were repeated twice. (C) Immunohistochemical staining of β-catenin in matched UF and MM tissues from UF patients (n = 9, magnification 20× and 65×) (D) Semi-quantitative intensity and frequency scores of cytoplasmic and nuclear β-catenin in the evaluated tissues as well as ratio presented as the mean ± SEM. At least 5 images represent different areas were used for statistical measurement. (E) Immunohistochemical staining for cyclin D1 and c-Myc, the products of β-catenin responsive genes, in matched UF and MM tissues from UF patients (n = 3, magnification 20× and 65×). (F) Cyclin D1 and c-Myc mRNA expression was measured by qRT-PCR analysis in matched UF and MM tissues from patients with UFs (n = 5). The mRNA levels were normalized to 18S rRNA and normalized values were used to generate the graph. Data are presented as the mean ± SEM of triplicate measurements for each patient. Experiments were repeated twice. *P < 0.05; **P < 0.001.

Next, we performed IHC staining to explore the expression and location pattern of β-catenin in 9 matched UF and MM tissues (Fig. 1C shows representative figures from 3 matched tissues). Interestingly, nuclear β-catenin, versus cytoplasmic, was significantly higher in UF compared with MM tissues. The nuclear:cytoplasmic ratio was 3.5-fold higher in UF compared with MM tissues (Fig. 1D, n = 9; P < 0.05). This data might explain that, although in some instances MM tissue may have higher expression of total β-catenin compared with UF tissue, nuclear β-catenin, and consequently its responsive genes, are still higher in UF. This was both on protein level (Fig. 1C, Fig. 1E) and gene expression level as shown in Fig. 1F, where gene expression of CCND1 and MYC were significantly higher in UF compared with MM tissues (n = 5; P = 0.0313 and P = 0.009 respectively).

Next, we confirmed our findings using the HuLM and UTSM cell lines (Fig. 2A and 2B) as well as primary cells isolated from matched UF and MM tissues (Fig. 2D and 2E). Using Western blot (WB) analysis, HuLM cells showed significant higher expression of total β-catenin and the products of its responsive genes, compared with UTSM, while p27 was lower (Fig. 2A). Nuclear and cytoplasmic fractionation of cells lysate showed similar tissue patterns, in which nuclear β-catenin was significantly higher in HuLM and lower in UTSM as compared with cytoplasmic fraction (P < 0.05) (Fig. 2B). More importantly, nuclear expression of β-catenin was significantly higher in HuLM compared with UTSM cells (P < 0.001). The qRT-PCR analysis showed significant higher expression of the β-catenin responsive genes CCND1, MYC, and AR as well as other markers that we identified from the literature, including HAS2, HAS3, OPCML, PITX2, MED12L, ISL1, and ANGPT2 (Fig. 2C). Similar findings of the higher expression of total and nuclear β-catenin in UF cells were also found in primary cells (n = 2) concomitant with increased expression of its responsive genes (Fig. 2D and 2E).

Figure 2.

Activation of β-catenin signaling in human uterine fibroids is associated with higher nuclear β-catenin protein expression. (A) Western blot analysis of protein lysates from uterine fibroid (HuLM) and myometrium (UTSM) cell lines for β-catenin, the products of its responsive genes cyclin D1, c-Myc, and androgen receptor (AR), and the product of the tumor suppressor gene p27. (B) Cytoplasmic (C) and nuclear (N) extracts were prepared from both UTSM and HuLM cells. A total of 30 µg each was analyzed by Western blots using anti–β-catenin antibody, poly (ADP-ribose) polymerase (PARP) (nuclear), and RhoGDI (cytoplasmic). Western blots were used to show the efficiency of the separation followed by graphical representation for the intensity of each protein band quantified and normalized with corresponding β-actin. Data are presented as the mean ± SD. (C) The mRNA expression of β-catenin responsive genes (CCND1, c-Myc, AR, HAS2, HAS3, OPCML, PITX2, MED12L, ISL1, and ANGPT2) were measured in HuLM and UTSM cells by qRT-PCR analysis. The mRNA levels were normalized to 18S rRNA and normalized values were used to generate the graph. Data are presented as the mean ± SEM of triplicate measurements. (D) Western blot analysis of protein lysates from primary cells isolated from uterine fibroid (F) and myometrium (M) (n = 2) for β-catenin and the protein products of its responsive genes cyclin D1, c-Myc, and androgen receptor (AR), and the product of the tumor suppressor gene p27. (E) Cytoplasmic (C) and nuclear (N) extracts were prepared from both M and F primary cells (n = 2). A total of 30 µg protein for each sample was loaded and analyzed by Western blots using anti–β-catenin, anti-PARP (nuclear), and anti-RhoGDI (cytoplasmic) antibodies, followed by graphical representation for the intensity of each β-catenin protein band quantified and normalized with corresponding β-actin. Data are presented as the mean ± SD. All experiments were repeated twice. *P < 0.05; **P < 0.001; ***P < 0.0001.

Mislocalization of β-catenin is causally linked to uterine fibroid phenotype

To explore whether nuclear accumulation of β-catenin in UF is contributing to its phenotype, we selected 3 β-catenin inhibitors (ICG-001, cordycepin, and XAV939) as well as a well-known β-catenin activator (Wnt3a) to examine their effects on HuLM and UTSM cells. First, we performed a MTT assay to examine their effect on HuLM cell proliferation. Cells were treated with gradient concentrations of 4 compounds for 24 and 48 hours and compared with untreated control. All β-catenin inhibitors showed significant growth inhibitory effect on HuLM cells in a dose- and time-dependent manner (37), while Wnt3a stimulated cell growth (37). Since ICG-001 showed the most powerful antiproliferative effect among the 3 tested inhibitors, it was selected for further experiments.

Next, we determined the effect of β-catenin inhibition and activation on HuLM and UTSM cells, respectively, at the molecular level. HuLM treatment with ICG-001 (12.5 µM) for 12 hours showed significant inhibition of β-catenin translocation from cytoplasm to nucleus by WB analysis upon lysate fractionation, while UTSM treatment with Wnt3a (20 ng/mL) for 8 hours showed increased β-catenin nuclear translocation (P < 0.05) (Fig. 3A). Inhibition of β-catenin nuclear translocation in response to ICG-001 treatment in HuLM cells for 24 hours was accompanied with significant decrease in the protein expression of its pro-growth responsive markers cyclin D1 and c-Myc as well as AR. Treatment of UTSM with Wnt3a for 24 hours resulted in increased protein expression of the former markers compared with untreated control (Fig. 3B). A similar finding regarding the inhibitory effect of ICG-001 treatment of HuLM on c-Myc and cyclin D1 was shown using immunofluorescence staining and confocal microscopy imaging (Fig. 3C).

Figure 3.

Effect of the β-catenin inhibitor ICG-001 or activator Wnt3a on β-catenin nuclear translocation and β-catenin responsive gene expression. (A) Cytoplasmic (C) and nuclear (N) extracts were prepared from HuLM cells treated with either ICG-001 (12.5 µM) or DMSO (control) for 12 hours and UTSM cells treated with either Wnt3a (20 ng/mL) or H₂O (control) for 8 hours. A total of 30 µg protein for each sample was loaded and analyzed by Western blots using anti–β-catenin, anti-PARP (nuclear), and anti-RhoGDI (cytoplasmic) antibodies, followed by graphical representation (lower panel) for the intensity of each β-catenin protein band quantified and normalized with corresponding β-actin. Data are presented as the mean ± SD. (B) Western blot analysis of HuLM and UTSM cells treated with either ICG-001 (12.5 µM) or Wnt3a (20 ng/mL), respectively, for 24 hours. A total of 30 µg each was analyzed by Western blots using anti–β-catenin and antibodies to c-Myc, cyclin D1, and AR, the proteins encoded by the β-catenin–responsive genes CCND1, c-Myc, and AR. (C) The effect of ICG-001 on protein expression of c-Myc and cyclin D1: 8 ×10⁴ HuLM cells were seeded on sterile glass cover slip in 6-well plates and treated with ICG-001 (12.5 µM) for 24 hours. C-Myc and cyclin D1 expressions were assessed using immunofluorescence staining and confocal laser microscopy (20×). (D and E) The effect of ICG-001 and Wnt3a on gene expression of β-catenin responsive genes. The mRNA expression of the CCND1, c-Myc, MED12L, ISL1, and ANGPT2 genes was measured in HuLM cells treated with ICG-001 (12.5 µM) and UTSM cells treated with Wnt3a (20 ng/mL) for 24 hours and compared with untreated control by qRT-PCR analysis. The mRNA levels were normalized to 18S rRNA and normalized values were used to generate the graph. Data are presented as the mean ± SEM of triplicate measurement. All experiments were repeated twice. *P < 0.05; **P < 0.001; ***P < 0.0001; NS, nonsignificant.

We also explored the effect of ICG-001 and Wnt3a on the gene expression of the same previously mentioned markers. The qRT-PCR analysis showed treatment of HuLM with ICG-001 for 24 hours significantly decreased the gene expression of CCND1, MYC, and ANGPT2 (P < 0.05, P < 0.0001 and P < 0.0001, respectively) while MED12L and ISL1 did not exhibit significant downregulation (Fig. 3D and 3E). In contrast, treatment of UTSM with Wnt3a for 24 hours showed significant upregulation of CCND1, MYC, ISL1, and ANGPT2 gene expression (P < 0.05), while MED12L was nonsignificantly increased (P = 0.056).

Next, we knocked down the CTNNB1 gene in HuLM cell using lentiviral particles to confirm the previous findings and avoid any possible off-target effects using the chemical inhibitors. WB analysis was used to confirm the knockdown efficiency (Fig. 4). Knockdown resulted in enormous downregulation of cyclin D1, c-Myc, and AR protein expression, confirming the important role of β-catenin in modulating downstream components including cyclin D1, c-Myc and AR in UF (Fig. 4).

Figure 4.

Effect of β-catenin genetic knockdown on its downstream signaling in UF cells. Cell lysates prepared from control (Scr.) and CTNNB1 knockdown cells using 2 different constructs of ShRNA lentiviral particles (Sh1 and Sh2) were analyzed by Western blot using anti-β-catenin, anti-cyclin D1, anti-c-Myc, and anti–AR antibodies. β-actin was used as the loading control.

Estrogen activates β-catenin signaling in human uterine fibroids via estrogen receptor

Several studies showed a crosstalk between estrogen and β-catenin in various diseases. Since UF is a hormone-dependent tumor, we wanted to explore whether estrogen has a direct effect on β-catenin signaling. First, we treated UTSM with range concentration of 17β-estradiol (E2) (10-100 nM) for 24 hours and measured β-catenin nuclear translocation by WB analysis on fractionated cell lysates. Results showed a significant increased β-catenin nuclear translocation in dose-dependent manner in response to E2 treatment (Fig. 5A; P < 0.05). Next, we measured the effect of E2 treatment on the RNA expression of β-catenin responsive genes in UTSM by qRT-PCR. E2 treatment significantly increased the RNA expression of CCND1, MYC, ISL1, and HAS3 (P < 0.05), while the expression of MED12L, ANGPT2, HAS2, and PITX2 did not reach significance in response to E2 treatment (Fig. 5B). Notably, we used UTSM cells in this experiment, since β-catenin signaling is less active in UTSM than in HuLM cells, so the effect of E2 treatment on β-catenin activation would be clearer.

Figure 5.

Estrogen and β-catenin crosstalk in human myometrium and uterine fibroid cells. (A) Cytoplasmic (C) and nuclear (N) extracts were prepared from UTSM cells treated with 17β estradiol (E2) (10 and 100 nM) or ethanol (as control) for 24 hours. A total of 30 µg protein for each sample was analyzed by Western blots using anti–β-catenin, anti-PARP (nuclear), and anti-RhoGDI (cytoplasmic) antibodies, followed by graphical representation for the intensity of each β-catenin protein band quantified and normalized with corresponding β-actin. Data are presented as the mean ± SD. (B) The mRNA expressions of β-catenin responsive genes (CCND1, c-Myc, MED12L, ISL1, ANGPT2, HAS2, HAS3, and PITX2) were measured in UTSM cells treated with either E2 (100 nM) or ethanol (as vehicle control) for 24 hours by qRT-PCR analysis. The mRNA levels were normalized to 18S rRNA and normalized values were used to generate the graph. Data are presented as the mean ± SEM of triplicate measurements. (C) The protein expression of β-catenin and its responsive markers (cyclin D1, c-Myc, and AR) was measured in HuLM cells treated with E2 (10 and 100 nM), E2 (100 nM) in combination with the estrogen receptor antagonist ICI182780 (100 nM) or ethanol (as vehicle control) for 24 hours by Western blot analysis. (D) The mRNA expression of β-catenin and its responsive genes (CCND1, c-Myc, AR, HAS2, HAS3, and PITX2) was measured in HuLM cells treated with either E2 (100 nM), E2 (100 nM) in combination with ICI182780 (100 nM) or ethanol (as control) for 24 hours by qRT-PCR analysis. The mRNA levels were normalized to 18S rRNA and normalized values were used to generate the graph. Data are presented as the mean ± SEM of triplicate measurement. (E) The mRNA expression of estrogen receptor (ER) was measured in HuLM and UTSM cells as well as HuLM cells treated with ICG-001 (12.5 µM) and UTSM cells treated with Wnt3a (20 ng/mL) for 24 hours and compared with untreated control by qRT-PCR analysis. The mRNA levels were normalized to 18S rRNA and normalized values were used to generate the graph. Data are presented as the mean ± SEM of triplicate measurements. All experiments were repeated twice. *P < 0.05; **P < 0.001; ***P < 0.0001; NS, nonsignificant.

Next, we investigated if E2 induced activation of β-catenin signaling is mediated through ER. Therefore, we treated HuLM cells with E2 in presence or absence of the ER antagonist ICI182780 (100 nM) for 24 hours. First, we observed increased protein expression of β-catenin as well as its responsive markers in response to treatment with increasing concentration of E2 by WB analysis (Fig. 5C) as well as RNA expression using qRT-PCR (Fig. 5D). Interestingly, blocking the ER receptor with ICI182780 abrogated the stimulatory effect of E2 on the β-catenin signaling pathway in UF cells both on the protein and gene levels (Fig. 5C and 5D). These data demonstrated that E2 stimulates β-catenin signaling through the ER.

Finally, we determined whether the β-catenin signaling modulates the estrogen expression using gain and loss of function approaches. We first measured the gene expression of ER in UTSM cells in response to β-catenin activator Wnt3a. Treatment with Wnt3a (20 ng/mL) for 24 hours increased ER expression by 3-fold compared with untreated control as determined by qRT-PCR (Fig. 5E; P < 0.05). HuLM treatment with β-catenin inhibitor ICG-001 (12.5 µM) for 24 hours significantly decreased ER gene expression by almost 5-fold (Fig. 5E; P < 0.0001). Obviously, we used each cell type specifically to measure ER gene expression in response to either β-catenin activation in UTSM or inhibition in HuLM cells, since ER expression is highly expressed in HuLM cells compared with UTSM (30-fold, Fig. 5E; P < 0.0001). These data showed that reciprocal interaction between estrogen and β-catenin contributes to the UF phenotype.

Uterine fibroids express higher levels of Class I HDAC enzymes than normal myometrial tissue

Although the role of HDACs have been characterized in many diseases (42–46), little is known about the role of HDACs and their interaction with other signaling pathways, such as β-catenin and estrogen in UFs. One study previously showed significantly higher expression of HDAC6 (a Class II HDAC member) in 24 UF samples compared with their matched MM tissues (47). Herein we explored the protein expression of Class I HDAC enzymes including HDAC1, 2, 3, and 8 in matched UF and MM (n = 6). Interestingly, the UF tissues showed higher expression of all measured HDACs compared with MM tissues in 5 patients (Fig. 6A). Moreover, we noted a clear correlation in the expression pattern of measured HDACs with β-catenin protein expression in the tested samples (Fig. 6A)

Figure 6.

(A) Expression of Class I HDAC enzymes in human uterine fibroid (F) and myometrium (M) tissues. Western blot analysis of protein lysates from matched F and M tissues (n = 6) for β-catenin, HDAC1, HDAC2, HDAC3, and HDAC8. (B) The effect of HDAC inhibitors (HDACis) on the proliferation of HuLM cells. 2 × 10³ HuLM cells were seeded in 96 well plates and treated with gradient concentration of the HDACis apicidin and HDACi VIII (0.5-20 µg/mL) for 24 and 48 hours. Cell proliferation was assessed in each time point by MTT assay. Individual data points are the mean ± SEM of triplicate measurements (as percentage of untreated control). *P < 0.05; **P < 0.001. The experiments were repeated twice. DMSO was used as vehicle control.

HDAC inhibitors show antiproliferative effect on uterine fibroid cells via inhibition of β-catenin signaling pathway

Following the finding of higher HDACs expression in UFs, from both current and previous studies (47), we tested the effect of 2 HDAC inhibitors on HuLM cells: apicidin, a Class I HDAC inhibitor, and HDACi VIII, a potent HDAC6 inhibitor (also known as J-001196 or CAY10603). First, we tested the antiproliferative effects of the two HDAC inhibitors on HuLM cells individually. Cells were treated with range concentrations (0.5–20 µg/mL) for 24 and 48 hours, and then cell growth was compared with untreated cultured cells, as determined by MTT assay (Fig. 6B). Both treatments showed a time- and concentration-dependent growth inhibitory effect, which was significant starting from concentration 5 µg/mL for apicidin after 24-hour treatment, while significant for the rest of tested concentrations of the 2 drugs at all time points (Fig. 6B; P < 0.05). Remarkably, HDACi VIII showed stronger antiproliferative effect than apicidin; therefore, we selected the concentrations of 10 µg/mL for apicidin and 5 µg/mL for HDACi VIII to use in the next set of experiments.

Next, to explore whether those HDAC inhibitors exert their anti-UF effect via inhibition of β-catenin signaling, we measured the nuclear and cytoplasmic expression of β-catenin in HuLM in response to 24 hours HDACis treatment and compared it with the untreated cells using WB analysis on fractionated lysates (Fig. 7A). Results showed that both treatments significantly decreased nuclear β-catenin expression compared with control, while not affecting the cytoplasmic fraction (Fig. 7A; P < 0.05). Consequently, protein expression of β-catenin responsive markers cyclin D1, c-Myc, and AR were decreased, as shown in Fig.7B. Using immunofluorescence staining and confocal microscopy imaging, we confirmed that cyclin D1 and c-Myc protein expressions were clearly decreased (Fig. 7C). Similar findings were observed on the gene expression level using qRT-PCR, where 24-hour treatment with HDAC inhibitors significantly decreased the gene expression of CTNNB1, CCND1, MYC, PITX2, and HAS2 (P < 0.05) (37), while HAS3 reduction was nonsignificant. Furthermore, both treatments greatly decreased the gene expression of ER as well as progesterone receptor (PR) (37).

Figure 7.

Effect of HDACis on β-catenin signaling in UF cells. (A) Effect of HDACis on nuclear expression of β-catenin in UF cells. Cytoplasmic (C) and nuclear (N) extracts were prepared from HuLM cells treated with apicidin (10 µg/mL) or HDACi VIII (5 µg/mL) for 24 hours. A total of 30 µg protein for each sample was loaded and analyzed by Western blots using anti–β-catenin, anti-PARP (nuclear), and anti-RhoGDI (cytoplasmic) antibodies, followed by graphical representation for the intensity of each β-catenin protein band quantified and normalized with corresponding β-actin. Data are presented as the mean ± SD. (B) The protein expression of β-catenin and its responsive markers (cyclin D1, c-Myc, and AR) were measured in HuLM cells treated with either apicidin (10 µg/mL) or HDACi VIII (5 µg/mL) or DMSO (as control) for 24 hours by Western blot analysis. (C) The effect of apicidin and HDACi VIII on protein expression of CyclinD1 and c-Myc measured by ICC. A total of 8x10⁴ HuLM cells were seeded on sterile glass cover slip in 6-well plates and treated with either apicidin (10 µg/mL) or HDACi VIII (5 µg/mL) or DMSO (as control) for 24 hours. The expression of c-Myc and cyclin D1 was assessed using immunofluorescence staining and confocal laser microscopy (40×). All experiments were repeated twice. *P < 0.05; **P < 0.001; ***P < 0.0001; NS, nonsignificant.

HDAC inhibitors induce cell cycle arrest in uterine fibroid cells

An increasing body of evidence shows that several HDAC inhibitors are capable of inducing cell cycle arrest and apoptosis in tumor cells (31). We sought to determine whether our 2 selected HDAC inhibitors would exhibit a suppressive effect on HuLM cells. Cells were incubated with apicidin (10 µg/mL) or HDACi VIII (5 µg/mL) for 24 hours, stained with propidium iodide, and analyzed using flow cytometry. As shown in Fig. 8A, both drugs decreased the percentage of the cell population in S-phase, compared with untreated control cells, while increasing the percentage of cells at G2/M phase, indicating that the HDAC inhibitors induced UF cell cycle arrest at the G2/M phase. Since cell cycle is heavily dependent on cyclin-dependent kinases (CDKs) and regulated by CDK inhibitor p21 (48), we first measured p21 protein expression in response to HuLM cells treatment with HDAC inhibitors by WB analysis. Both treatments induced a significant increase in p21 protein expression compared with untreated control (Fig. 8B). In addition, we measured p21 gene expression, as well as p27 and CDK1, using qRT-PCR. Results showed that both treatments significantly increased p21 and p27 and decreased CDK1 gene expression compared with untreated control (Fig. 8C; P < 0.05).

Figure 8.

Effect of HDACis on cell cycle– and apoptosis-related markers in UF cells. (A) Cell cycle assay of HuLM cells stained with propidium iodide was analyzed using flow cytometry 24 hours after treatment with apicidin (10 µg/mL), HDACi VIII (5 µg/mL) or DMSO (as control). Left, G1 phase; Middle, S phase; Right, G2/M phase followed by Quantitative analysis of cell percentage at each phase in the 3 indicated groups. (B) The protein expression of p21 was measured in HuLM cells treated with either apicidin (10 µg/mL) or HDACi VIII (5 µg/mL) for 24 hours by Western blot analysis. (C) The mRNA expression of cell cycle–related genes (CDK1, p21 and p27) were measured in HuLM cells treated with either apicidin (10 µg/mL) or HDACi VIII (5 µg/mL) for 24 hours by qRT-PCR analysis. The mRNA levels were normalized to 18S rRNA and normalized values were used to generate the graph. Data are presented as the mean ± SEM of triplicate measurement. (D) The protein expression levels of apoptosis related markers (total PARP, cleaved PARP, Bcl₂, caspase 3, and active caspase 3) were measured in HuLM cells treated with either apicidin (10 µg/mL) or HDACi VIII (5 µg/mL) for 24 hours by Western blot analysis. (E) Apoptosis assay of HuLM cells treated with apicidin (10 µg/mL) or HDACi VIII (5 µg/mL) for 24 hours subjected to flow cytometry analysis after Annexin V-APC and 7-AAD dual staining, followed by quantitative analysis of cell percentage at apoptosis and necrosis stages in the 3 indicated groups.

HDAC inhibitors induce apoptosis in uterine fibroid cells

HDAC inhibitors have been shown to induce apoptosis in tumor cells by regulation of pro-apoptotic and anti-apoptotic genes (31, 49). To test the apoptosis-inducing ability of apicidin and HDACi VIII in UFs, HuLM cells were treated with 10 µg/mL and 5 µg/mL, respectively, for 24 hours, and the protein expression of key apoptosis markers was measured by WB analysis. Both treatments significantly increased the formation of cleaved poly (ADP-ribose) polymerase (PARP) and active caspase 3 compared with untreated control (P < 0.05), while total PARP and caspase 3 were decreased (Fig. 8D). Caspase-3 is a critical executioner of apoptosis as it cleaves many proteins like PARP, which is implicated in apoptosis; this cleavage of PARP is considered a hallmark event in apoptosis. Moreover, both HDAC inhibitor treatments decreased the protein expression of anti-apoptotic protein Bcl₂ compared with control (Fig. 8D). In addition, we performed a cell apoptosis assay in which the treated cells were stained with Annexin V-APC and 7-AAD, and analyzed by flow cytometry. The results showed that apicidin and HDACi VIII induced apoptosis (4.89- and 8.14-fold, respectively) as well as necrosis in HuLM cells (2.69- and 3.42-fold respectively) (Fig. 8E). Overall, HDAC inhibition showed a potent apoptosis-inducing ability in UF cells.

Discussion

UFs are the most important benign neoplastic threat to women’s health worldwide, with annual health care costs estimated in the hundreds of billions of dollars. Unfortunately, there are limited treatment options available, especially for safe long-term treatment. Deep understanding of the mechanisms underlying UF etiology will help to develop new therapies for UFs.

The literature contains frequent reports of β-catenin dysregulation in UFs. However, there is some discrepancy concerning differential β-catenin expression and location in UF and normal MM (18, 19). In the present work, nuclear β-catenin expression was significantly higher in UF tissues compared with their matched MM tissues. Similar findings for nuclear β-catenin were observed in a UF cell line and isolated primary cells compared with normal MM cells, whereas the cytoplasmic fraction was higher in MM tissues and inconsistent in primary cells. These results can explain the inconsistent differential expression of total β-catenin that has been shown previously, as well as in the current study. Interestingly, expression of β-catenin downstream targets cyclin D1 and c-Myc was higher in UF tissues at both protein and RNA levels including the samples that showed higher total β-catenin in MM than in UFs. These data demonstrate for the first time that mislocalization of β-catenin is linked to its activated signaling in human UFs. Previously, our group reported that UFs with MED12 mutations showed higher expression of total β-catenin compared with their matched MM tissues in 5 patients by WB analysis (10). Other investigators demonstrated a paracrine role for the WNT/β-catenin pathway that enables mature MM or UF cells to send mitogenic signals to neighboring tissue stem cells in response to estrogen and progesterone, leading to the growth of UFs (16). Proteomic analysis as well as qRT-PCR showed higher expression of β-catenin in UFs compared with adjacent normal MM in another study (17). Zaitseva et al showed elevated β-catenin gene expression in UF compared with matched MM tissues (in 10 patients) collected during the proliferative phase while tissues in secretory phase did not show significant difference (18). However, the opposite pattern of β-catenin protein expression was observed in the same study using WB analysis, where differential expression was significant in secretory phase tissues and not significant in proliferative phase tissues (18). Immunostaining of β-catenin in the same study revealed weak to moderate intensity in both MM and UF tissue with stronger staining around blood vessels. Regarding location, the authors mentioned diffuse staining through the cell membrane and cytoplasm, with minimal nuclear staining (18). Moreover, another study showed that β-catenin was not differentially expressed in UF and MM tissues (12 patients) obtained from different stages of the menstrual cycle at both the protein and RNA expression levels (19). Importantly, none of these studies explored the differential expression of β-catenin responsive proliferative markers, such as cyclin D1 and c-Myc, or how these markers correlated with β-catenin expression and translocation in UFs. Our study showed that β-catenin inhibition in UF cells, using different β-catenin chemical inhibitors, decreased UF cell growth and proliferation, while activation of β-catenin signaling promoted its growth. At the molecular level, we observed that this effect was correlated with nuclear β-catenin expression level and subsequent transcription and translation of cyclin D1 and c-Myc in both cell lines and primary isolated cells. Genetic knockdown of CTNNB1 further demonstrated marked downregulation of cyclin D1 and c-Myc, which was consistent with the study using the chemical inhibition approach. Previously, a study showed UF cell growth inhibition in response to 3 β-catenin inhibitors including XAV939, which we used in our study, along with decreased Wnt/β-catenin pathway activation using a TCF reporter construct (50). In our study, we used 2 additional β-catenin inhibitors that have shown efficacy in other diseases (51, 52) and that were never tested before in UFs. We also measured AR expression along with cyclin D1 and c-Myc in all our experiments. β-catenin was identified to be a co-activator of the AR (53, 54). In a prostate cancer study, β-catenin increased AR transcriptional activation by androstenedione and estradiol. Moreover, co-immunoprecipitation of β-catenin with AR showed that the 2 molecules are present in the same complex (53). Another study concluded that beta-catenin augments the ligand-dependent activity of AR in prostate cancer cells (54). Although UF is known to be estrogen- and progesterone-dependent, previous studies demonstrated that the AR exerted proliferative and anti-apoptotic functions in MM cells and that IHC analysis showed it was highly expressed in UF compared with matched MM tissues (14 patients) (41). Our data confirmed the higher protein expression of AR in UF tissues and cells. In addition, its expression was decreased upon β-catenin chemical inhibition or genetic knockdown and increased after activation. These data suggested that AR is a novel downstream target of β-catenin in UFs.

An increasing number of nuclear receptors have been shown to interact with β-catenin and cause alterations in cell proliferation and tumorigenesis (55). Many studies have revealed crosstalk between estrogen and β-catenin signaling at physiological levels seen in brain (56) and uterus (24), as well as in hormone-dependent tumors such as breast cancer (26) and endometriosis (23, 27). Surprisingly, the important interaction between ER and β-catenin signaling has not been well identified in UFs. A study showed that β-catenin gene expression was higher in UF tissues at the proliferative phase compared with the secretory phase (18), indicating an essential role for estrogen in regulating β-catenin expression. In this study, we demonstrated that 17β-estradiol (E2) stimulated nuclear translocation of β-catenin in MM cells in a dose-dependent manner, therefore enhancing the transcription of downstream proliferative genes such as CCND1 and MYC. We used E2 in this study since it is the most potent estrogen produced in the body (57). Moreover, treatment of UF cells with ICI182780 to antagonize ER function abolished the E2-induced expression of β-catenin and its downstream targets. These data are consistent with previous findings in human endometrial stromal cells (23) and in breast cancer cells (26), demonstrating that estrogen is linked with β-catenin signaling. Interestingly, E2 and β-catenin showed reciprocal interaction, since β-catenin inhibition or activation resulted in ER enhanced or diminished expression, respectively. Varea et al showed that E2 activates the phosphoinositide 3-kinase (PI3K)/Akt signaling pathway and inactivates GSK3β, thus increasing β-catenin activity and nuclear translocation in neuron cells (58). Zhang et al showed that E2 promotes ER binding to the estrogen receptor element (ERE) site within the CTNNB1 promotor (27). This can explain how E2 treatment increased both β-catenin gene and protein expression in UF cells, since estrogen can activate β-catenin in different ways. However, further research is required to explore the exact molecular mechanism underlying E2 and β-catenin regulation.

HDACs are a class of enzymes that remove acetyl groups from an N-acetyl lysine amino acid on a histone, allowing the histones to wrap the DNA more tightly, thereby regulating gene expression. In addition, nonhistone proteins are also targeted by HATs and HDACs, modulating various biological events. Studies have demonstrated that abnormal HDAC signaling contributes to many diseases, including tumorigenesis (30, 31). However, little is known about the role of HDACs in UF pathogenesis. HDACs are classified into 4 classes based on their sequence homology to the yeast original enzymes and domain organization (59). Wei et al previously showed that HDAC6 (a Class II HDAC member), localized in the cytoplasm, was highly expressed in UF compared with in normal MM (52). In addition, Wei et al showed a close correlation between HDAC6 and ER expression in UF tissues (47).

Givin the fact that HDAC Class I members play an important roles in many diseases, and have not yet been investigated in UFs, we measured the expression levels of Class I HDAC enzymes (HDAC1, 2, 3, and 8), and found that the expression levels of Class I HDACs were upregulated in UF tissues.

Importantly, β-catenin has been shown to be a nonhistone substrate of HDACs, and acetylation status of β-catenin is tightly linked with its signaling activation. Based on previous studies by our group and others, we selected the Class I HDAC inhibitor apicidin and the HDAC6 inhibitor HDACi VIII for exploration of their anti-UF effect. Treatments with both HDAC inhibitors effectively inhibited UF cell proliferation while inducing cell cycle arrest and apoptosis. At the molecular level, both treatments decreased nuclear β-catenin expression and consequently, cyclin D1, c-Myc, and AR expression, suggesting that inhibition of Class I and II HDAC members suppressed the β-catenin signaling. This study is consistent with a study by Li et al showing the key role of HDAC6 in nuclear translocation of β-catenin and its signaling in colon cancer cells (60). One possible mechanism was HDAC6 deacetylation of β-catenin at Lys-49, thus reducing its phosphorylation and inducing its nuclear accumulation, which is repressed by using HDAC6 inhibitor (60). For the role of Class I HDAC members in β-catenin signaling, a previous study showed that HDAC3 increased β-catenin activity via the AKT/GSK3β pathway in breast cancer, so selective HDAC3 inhibition showed in vitro and in vivo efficacy against breast cancer via inhibition of β-catenin (61). Wang et al explored the in vitro and in vivo efficacy of the Class I HDAC inhibitor JSL-1 in uveal melanoma (62) and demonstrated that JSL-1 was capable of decreasing the protein levels of both total and nuclear β-catenin as well as c-Myc and cyclin D1. A possible mechanism was that JSL-1 accelerated β-catenin turnover in a lysosome-dependent pathway. Our studies clearly demonstrated that activated β-catenin signaling in UFs could be suppressed by HDAC inhibition; however, the detailed mechanism of HDAC-mediated regulation of β-catenin in UF needs further investigation.

In this study, treatments with Class I and II HDAC inhibitors induced cell cycle arrest and apoptosis via increasing expression of several related markers, such as p21, cleaved PARP, and active caspase 3. This finding is consistent with previous studies in other types of tumors using several HDAC inhibitors (31, 62). Chen et al recently showed that the natural compound resveratrol, known as a pan-HDAC inhibitor (63), inhibited UF growth via inhibition of β-catenin and induction of apoptosis and cell cycle arrest (64).

Interestingly, our 2 tested HDAC inhibitors decreased ER and PR transcription in UFs, confirming previous studies showing that HDACs play an essential role in regulation of estrogen signaling (35). HDAC6 inhibition by siRNA caused substantial reduction of ER expression in Eker leiomyoma tumor cells (ELT3) (47). Sulaiman et al showed that co-suppression of β-catenin, HDAC and ER using clinically relevant low-dose inhibitors effectively repressed breast cancer cells (34). Figure 9 shows a proposed model displaying activated β-catenin signaling as well as its crosstalk with estrogen and HDACs in UF pathogenesis and summarizes the main findings in this study.

Figure 9.

Proposed model displaying activated β-Catenin signaling, its crosstalk with estrogen and histone deacetylase (HDAC) in uterine fibroid (UF) pathogenesis. β-catenin signaling is activated in uterine fibroids and modulated by Wnt, estrogen, and HDAC signaling pathways. Wnt3a and 17β-estradiol (E2) stimulate β-catenin translocation, therefore enhance the transcription of CCND1, c-Myc, and the androgen receptor gene (AR), which are the β-catenin–regulated genes. The activated β-catenin signaling is inhibited by β-catenin inhibitor (ICG-001), ER antagonist (ICI-182780), and HDAC inhibitors (apicidin and HDACi VIII), leading to decreased nuclear translocation of β-catenin and blocking β-catenin downstream signaling, therefore suppressing the phenotype of UFs.

In conclusion, our studies demonstrate for the first time that β-catenin signaling is activated in UFs with evidence of mislocalization of β-catenin and increased expression of β-catenin downstream components. The activated β-catenin signaling is tightly linked to estrogen and HDAC activities in UFs. These novel findings highlight the importance of β-catenin signaling in UF pathogenesis, and provide signaling targeting options for suppressing the UF phenotype.

Glossary

Abbreviations

CDK

cyclin-dependent kinase

CK-Iα

casein kinase-Iα

DMSO

dimethyl sulfoxide

E2

17β-estradiol

ER

estrogen receptor

GSK3β

glycogen synthase kinase 3β

HAT

histone acetyltransferase

HDAC

histone deacetylase

HDACi VIII

histone deacetylase inhibitor VIII

HuLM

immortalized human uterine fibroid cell line

IHC

immunohistochemistry

MED12

mediator complex subunit 12 gene

MM

myometrium

PARP

poly (ADP-ribose) polymerase

PI3K

phosphoinositide 3-kinase

qRT-PCR

quantitative real-time polymerase chain reaction

UF

uterine

UTSM

immortalized human uterine smooth muscle cells

Additional Information

Reprint requests should be addressed to the corresponding author.

Disclosure Summary: Ayman Al-Hendy has been a consultant and participated in advisory boards for Allergan plc, Bayer, Repros, Myovant, MD Stem Cells, AstraZeneca, Wyeth and AbbVie.

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