Histone Deacetylase Inhibitors Induce Differentiation of Human ...

0013-7227/05/$15.00/0 Printed in U.S.A.

Endocrinology 146(12):5365–5373 Copyright © 2005 by The Endocrine Society doi: 10.1210/en.2005-0359

Histone Deacetylase Inhibitors Induce Differentiation of Human Endometrial Adenocarcinoma Cells through Up-Regulation of Glycodelin Hiroshi Uchida, Tetsuo Maruyama, Takashi Nagashima, Hironori Asada, and Yasunori Yoshimura Department of Obstetrics and Gynecology, Keio University School of Medicine, Shinjuku, Tokyo 160-8582, Japan Histone reversible acetylation, which is controlled by histone acetyltransferases and deacetylases, plays a fundamental role in gene transcription. Histone deacetylase inhibitors (HDACIs), such as trichostatin A (TSA) and suberoylanilide hydroxamic acid (SAHA), have been characterized not only as anticancer drugs, but also as cytodifferentiation-inducing agents. In human endometrium, postovulatory production of progesterone directs estrogen-primed endometrial glandular cells to differentiate and thereby produce a number of unique bioactive substances, including glycodelin, that are critical for implantation at the secretory phase of the menstrual cycle. In this study, we show that TSA and SAHA, belonging to the hydroxamic acid group of HDACIs, can induce the phenotype of a human endometrial adenocarcinoma cell line, Ishikawa (originally derived from the glandular component of the endometrium), to differentiate to closely resemble normal en-

H

UMAN UTERINE ENDOMETRIUM undergoes proliferation, differentiation, tissue breakdown, and shedding during the menstrual cycle under the influence of estrogens and progesterone (P4). Coordinate and sequential production of these ovarian steroid hormones is required for proper and cyclical changes in human endometrium, whereas their aberrant production predisposes to the development of endometrium-derived diseases (1). For instance, prolonged exposure to unopposed estrogen without progesterone is thought to give rise to endometrial hyperplasia and carcinoma. Progestin-based therapies are widely used for endometrium-derived disorders, including well-differentiated endometrial cancer, endometriosis, and endometrial dysfunction due to an insufficient responsiveness to progestins (2, 3). The basic conceptual mechanism underlying the treatment with progestin is to counteract the proliferating effect of estrogens and induce an appropriate differentiation of the disordered endometrium. First Published Online August 25, 2005 Abbreviations: E2, 17␤-Estradiol; EGFP, enhanced green fluorescent protein; E2P4, 17␤-estradiol and progesterone; GAPDH, glyceraldehyde3-phosphate dehydrogenase; Gd1-EGFP, enhanced green fluorescent protein-fused glycodelin; HAT, histone acetyl transferase; HC toxin, Helminthosporium carbonum toxin; HDAC, histone deacetylase; HDACI, histone deacetylase inhibitor; LIF, leukemia inhibitory factor; P4, progesterone; PR, progesterone receptor; SAHA, suberoylanilide hydroxamic acid; SBHA, suberoyl bis-hydroxamic acid; siRNA, small interference RNA; Sp1, specificity protein 1; TSA, trichostatin A. Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the endocrine community.

dometrial epithelium in a time- and dose-dependent manner, as determined by morphological changes, synthesis of glycogen, and expression of secretory phase-specific proteins, including glycodelin. The proliferation- and differentiationmodulating effects elicited by TSA and SAHA at their optimal concentrations were comparable or more potent than those exerted by combined treatment with progesterone and estradiol. Furthermore, the gene silencing of glycodelin by small interference RNA resulted in the blockade of HDACI-induced differentiation in Ishikawa cells, suggesting the requirement for glycodelin for endometrial epithelial differentiation. Our results collectively indicate that TSA and SAHA are potent differentiation inducers for endometrial glandular cells, providing a clue for a possible therapeutic strategy to modulate endometrial function by targeting glycodelin. (Endocrinology 146: 5365–5373, 2005)

Ovarian steroid hormones, including P4, act mainly via their corresponding nuclear receptors and hormone-responsive elements in assistance with coactivators and corepressors, regulating the transcriptional activity of the target genes (4), although nongenomic actions of steroid hormones have recently emerged (5). It is well known that several coactivators are histone acetyltransferases (HATs), whereas corepressors mediate transcriptional repression through the recruitment of the histone deacetylase (HDAC) complex (4). HATs and HDACs coordinately control the level and pattern of histone acetylation, modify the local chromatin structure, and thereby regulate gene transcription, which eventually affects various biological processes, such as cell growth, differentiation, and apoptosis (4). Many HATs and HDACs have been identified in the past decade, coinciding with the discovery of specific HDAC inhibitors (HDACIs), such as trichostatin A (TSA) (6). These inhibitors increase the level of histone acetylation in many cell types, thereby inducing the expression of specific preprogrammed genes whose products, in turn, lead to cell growth arrest, differentiation, and apoptotic cell death (7). We have previously reported that TSA, a potent and specific HDACI, potentiates the P4-induced differentiation of stromal cells (decidualization) isolated from human cycling endometrium (8). Although stromal cells are one of the major components of human endometrium, luminal and glandular epithelial cells, other major components, are also important because they are located in the front line of the endometrium and therefore have the unique potential to directly interact with the embryo, producing many bioactive substances, in-

5365

5366

Endocrinology, December 2005, 146(12):5365–5373

Uchida et al. • Glycodelin in Endometrial Differentiation

cluding glycodelin (9), that are critical for implantation and placentation. Furthermore, many endometrial neoplasms are originated from the luminal and glandular surface epithelium. Recently, HDACIs have been reported to profoundly inhibit cell proliferation in endometrial cancer cells (10). These collectively prompted us to examine whether various HDACIs, including TSA, have the potential to induce and/or enhance the differentiation of endometrial epithelial cells. We here demonstrate that the hydroxamic acid group of HDACIs alone can induce functional and morphological differentiation in a human endometrial glandular cell line, Ishikawa, through up-regulation of glycodelin. The differentiation-inducing effects of TSA and suberoylanilide hydroxamic acid (SAHA) are comparable or more potent than those of combined treatment with P4 and estrogen. Our data suggest that HDACIs may be useful in the investigation of the molecular mechanisms underlying steroid-induced endometrial cell differentiation and that they may have therapeutic potential for endometrium-derived diseases, such as endometrial cancer and endometriosis. Materials and Methods

cultured with or without 10 nm 17␤-estradiol (E2) plus 1 ␮m P4 in combination with the indicated concentrations of HDACIs for different time periods according to the experimental procedures. Cell images were acquired using a Leica TCS SP2 confocal microscopy system with a Leica DMIRE2 inverted microscope (Leica Microsystems, Heidelberg, Germany).

Plasmids Genomic DNA was prepared from Ishikawa cells using a DNeasy tissue kit (Qiagen, Hilden, Germany), and the promoter region (⫺304 to ⫹20 bp) of glycodelin gene (12) was amplified by PCR using primers, as shown in Table 1. To construct a reporter plasmid pGL3 Gd 304, the PvuII/SacI PCR fragment was subcloned into the blunted KpnI-SacI site of pGL3 plasmid (Promega Corp., Madison, WI). Total RNA was extracted from cultured Ishikawa cells using an RNeasy mini kit (Qiagen). Glycodelin cDNA without signal sequence and stop codon was amplified by RT-PCR using the primers shown in Table 1. The forward primer has an EcoRI site and Kozak sequence, and the reverse primer has a BamHI site, but not a stop codon. The PCR product and pEGFP-N3 plasmid (BD Clontech) were both digested with EcoRI and BamHI and ligated to generate an expression plasmid harboring EGFP-fused glycodelin gene (pcGd1-EGFP). DNA sequences were verified using an automated Applied Biosystems sequencer (Foster City, CA) and the BigDye Terminator Kit (PerkinElmer, Boston, MA).

RT-PCR

Reagents Phenol red-free MEM and fetal bovine serum were purchased from Invitrogen Life Technologies (Tokyo, Japan). TSA was obtained from Wako Biochemicals (Osaka, Japan). Other HDACIs, such as HC (Helminthosporium carbonum) toxin, Scriptaid, Nullscript, SAHA, and suberoyl bis-hydroxamic acid (SBHA), were purchased from BIOMOL (Plymouth Meeting, PA). All oligonucleotides were synthesized by Invitrogen Life Technologies. Antibodies against glycodelin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), leukemia inhibitory factor (LIF; Santa Cruz Biotechnology, Inc.), MAPK (Upstate Biotechnology, Inc., Lake Placid, NY), enhanced green fluorescent protein (EGFP; BD Clontech, Palo Alto, CA), and horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA) were purchased from commercial sources. Unless indicated otherwise, all other chemicals were obtained from Sigma-Aldrich Corp. (St. Louis, MO) and Wako Biochemicals.

Cell culture The human endometrial adenocarcinoma cell line Ishikawa [clone 3-H-12, positively expressing estrogen receptor ␣ and progesterone receptor (PR)] (11) was provided by Dr. M. Nishida (National Kasumigaura Hospital, Ibaragi, Japan). Cells were maintained in phenol-red free MEM supplemented with 10% charcoal-treated fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin at 37 C under 5% CO2 in a humidified incubator and were used within 10 passages. Cells were

Total RNA was extracted from cultured Ishikawa cells using an RNeasy mini kit (Qiagen). RT-PCR was carried out with 200 ng total cellular RNA using the One-Step RT-PCR kit (Qiagen). Primers used to amplify human glyceraldehyde-3-phosphate dehydrogenase (GAPDH), glycodelin (13), and LIF (14) are shown in Table 1. After PCR amplification, samples were electrophoresed in 3% agarose gels, followed by photographic recording of the ethidium bromide-stained gels by FAS-III MINI (Toyobo, Tokyo, Japan). The band intensities were measured using the public domain National Institutes of Health Image program, version 1.62. The relative ratio was calculated as the densitometry of glycodelin or LIF divided by that of GAPDH, and the relative ratio of untreated cells was set at 1.0. All experimental data for RT-PCR represent the results obtained from three independent experiments.

Immunoprecipitation and immunoblotting Semiconfluent Ishikawa cells cultured in the indicated medium for various periods were lysed on ice with RIPA buffer [20 mm Tris-HCl (pH 7.5), 150 mm NaCl, 1 mm EDTA, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 1 mm Na3VO4, 50 mm sodium fluoride, and 1 mm Na2MoO4) containing protease inhibitor cocktail (Roche, Basel, Switzerland). Protein concentrations were determined using the DC protein assay kit (Bio-Rad Laboratories, Hercules, CA) with BSA as a standard. Each 250 ␮g protein was immunoprecipitated by incubation with 4 –5 ␮l antiglycodelin or anti-LIF polyclonal antibody and protein G-Sepharose beads (Amersham Biosciences, Piscataway, NJ) for 3 h at 4 C. The

TABLE 1. Human oligonucleotide primers for amplification of the glycodelin promoter and coding region and for RT-PCR in Ishikawa cells Gene

Glycodelin promoter Glycodelin cDNA GAPDH Glycodelin LIF a b

Positiona

Sequence (5⬘–3⬘)

AGC CCT CGG CGG TCA CTG AAG ACG TGG TGT

AGT GGC AAT GAT CCA CTT TTG GCA TTC AAT

TCG TGT TCG CCG TCT CAC GCA CGG TGC AGA

AGA CAG CCA AAA TCC CAC GGG CTC ACT GAA

CCA AAA TGG CGG AGG CTT ACC TTC GGA TAA

GCT ATG ACA CAC AGC CTT TGG CAT AAC AGA

TG CC TCC GGC G GA CAC CTG ATG GGG

⫺1470–⫺1451 ⫹796–⫹777b 55–74 540–521 221–239 792–773 94–116 537–515 38–57 202–176

b

CCC AGA CCA A TCT TCC A

TC TT CAT TGG

GenBank accession no.

Ref.

M34046

12

M34046 J04038 M34046

13

M63420

14

Numbers in column of position represents nucleotide location in the coding region sequence, except glycodelin promoter. Numbers in glycodelin promoter represent nucleotide location in glycodelin genome sequence.


immune complexes were washed four times with RIPA buffer and then resuspended in 2⫻ sodium dodecyl sulfate buffer. Samples were separated by electrophoresis on an 8% or 15% SDS-PAGE gel and transferred onto a polyvinylidene difluoride membrane. After incubation with antiglycodelin or anti-LIF antibody, followed by appropriate horseradish peroxidase-conjugated secondary antibody, the immunoreactive proteins were detected by the enhanced chemiluminescence method (Amersham Biosciences). As an internal control, the expression of MAPK in 25 ␮g of each input cell lysate was visualized using immunoblotting with anti-MAPK antibody. All experimental data for immunoprecipitation and immunoblotting represent the results obtained from three independent experiments.

Promoter assay Ishikawa cells (4 ⫻ 104) were plated onto 24-well plates 1 day before transient transfection with pGL3 alone or pGL3 Gd 304 plasmid using FuGene6 (Roche). After transfection, Ishikawa cells were cultured for 3 d in medium containing E2, E2 plus P4 (E2P4), or the various concentrations of SAHA and TSA, followed by luciferase assay using a TD-20/20 luminometer (Turner Designs, Sunnyvale, CA) and a dual luciferase assay kit (Promega Corp.). Luciferase activity was represented as the mean and sem of the relative light units obtained from three independent experiments. The relative luciferase activity of pGL3 Gd 304-transfected Ishikawa cells in the control vehicle was set at 1.0.

Measurement of the intracellular glycogen content The glycogen content of cultured Ishikawa cells was measured as described previously (15). In brief, Ishikawa cells grown in 100-mm culture dishes were harvested, collected by centrifugation, and resuspended in 150 ␮l KOH, followed by boiling at 100 C for 20 min. Twentyfive microliters of saturated sodium sulfate solution and 200 ␮l 95% ethanol were added to the samples and boiled again. The pellet obtained by centrifugation at 3000 ⫻ g for 5 min at 4 C was dissolved with 1 ml distilled water, mixed with 5 vol anthrone solution (7.6 ml distilled water, 19.6 ml sulfic acid, 25 mg thiourea, and 25 mg anthrone), and finally boiled for 15 min. After chilling, the glycogen concentration was measured by luminometer at 620 nm, with rat liver glycogen as a standard. The glycogen ratio was calculated as the glycogen concentration divided by the intracellular protein concentration. The relative glycogen ratio of the untreated cells was set at 1.0.

Cell growth Ishikawa cells (4 ⫻ 104) were plated onto 24-well plates (d 0) and grown for 12 d in the indicated medium, with refreshing every 3 d. Cells were harvested by trypsinization, and the number of viable cells was counted every 3 d.

Small interference RNA (siRNA) and its transfection siRNA-targeting GAPDH was obtained from Ambion (Austin, TX). The target sequence in the glycodelin gene was 5⬘-CGTGGCCCTGGTCTGTGGTGT-3⬘. Sense RNA (5⬘-UGGCCCUGGUCUGUGGUGUUU-3⬘) and antisense RNA (5⬘-ACACCACAGACCAGGGCCACG3⬘) were independently synthesized, annealed, and then used as glycodelin siRNA (Qiagen). Either GAPDH or glycodelin siRNA was transfected into Ishikawa cells using RNAiFect transfection reagent (Qiagen).

Statistical analysis All experimental data from the bioassays represent the results obtained from three independent experiments of each duplicate or triplicate assay and are expressed as the mean ⫾ sem. Statistical analysis was performed using ANOVA, followed by a Bonferroni test.


5367

Results The hydroxamic acid group of HDACIs, such as TSA, SAHA, and SBHA, up-regulates differentiation-associated genes in endometrial adenocarcinoma cells (Ishikawa)

We previously reported that TSA alone can induce differentiation in stromal cells isolated from human cycling endometrium (8). These results prompted us to examine the possible differentiation-inducing effect of TSA on endometrial glandular cells, another major cell component of human endometrium. To address this question, we have employed a human well-differentiated adenocarcinoma cell line called Ishikawa (clone 3-H-12) instead of primary culture of endometrial glandular cells, because the primary culture, in particular, long-term culture, is technically difficult, and the availability of human endometrial samples is limited. Ishikawa cells harbor estrogen receptor ␣ and PR; therefore, they have been widely used for studies on the physiology and pathophysiology of normal human endometrial glandular cells and well-differentiated endometrial adenocarcinoma (11). As shown in Fig. 1A, treatment with E2P4 increased the mRNA expression of glycodelin, an established differentiation marker for endometrial glandular cells (9), in Ishikawa cells in a time-dependent manner. Its expression attained at the maximal level after 9 d of treatment. The in vitro kinetics appear to be consistent with the in vivo expression pattern in which glycodelin is prominently expressed in the glandular and luminal epithelium at the P4-dominated late secretory phase endometrium (9). TSA alone up-regulated glycodelin mRNA more promptly than ovarian steroids (Fig. 1A). The expression levels of GAPDH were constant throughout treatment with E2P4 and TSA (Fig. 1A). We then tested the effects of other HDACIs on the glycodelin mRNA expression in Ishikawa cells cotreated without or with E2P4. Similar to TSA, treatment with SAHA or SBHA, each belonging to the hydroxamic acid group of HDACIs, induced mRNA expression of glycodelin in a dosedependent fashion (Fig. 1B). In contrast, HC toxin, a nonhydroxamic acid-type HDACI, did not affect the glycodelin expression (Fig. 1B). Although Scriptaid is a hydroxamic acid-type HDACI, it could not elicit the apparent expression of glycodelin gene, similar to Nullscript, that is an inactive compound of Scriptaid (Fig. 1B). Although 500 ␮m TSA synergistically augmented the induction of glycodelin mRNA expression by E2P4, the other concentrations of TSA and other HDACIs tested did not show any significant enhancing effects in Ishikawa cells (Fig. 1B). SAHA and TSA induce protein expression of glycodelin

In addition to the mRNA level, we examined the protein level of glycodelin in Ishikawa cells treated without or with ovarian steroids, SAHA, or TSA. To address the specificity of the antiglycodelin antibody used in this study, we first tested whether it can recognize the exogenously overexpressed Gd1-EGFP. As shown in Fig. 2A, antiglycodelin antibody specifically reacted with a band corresponding to the Gd1EGFP (black arrow in the left panel), but not EGFP alone, whereas anti-EGFP-antibody recognized both EGFP and Gd1-EGFP (white and black arrows, respectively, in the right

5368


FIG. 1. HDACIs and ovarian steroid hormones induce mRNA expression of glycodelin in Ishikawa cells. Ishikawa cells were treated with 10 nM E2 plus 1 ␮M P4 (E2P4) or 250 nM TSA for the indicated times (A). The cells were cultured for 3 d in medium containing E2P4 or the indicated concentrations of various HDACIs in combination with or without E2P4 (B). Total RNA was extracted from the treated cells and subjected to RT-PCR analysis of GAPDH and glycodelin (A and B). In each RT-PCR analysis, a single representative ethidium bromide staining of agarose gel is shown. Bars indicate the mean and SEM relative ratio obtained from three independent experiments, as described in Materials and Methods. Asterisks show significant differences compared with the nonstimulated Ishikawa cells (P ⬍ 0.05).

panel). Figure 2B showed that E2 alone induced the expression of glycodelin protein in a time-dependent manner, which was dramatically enhanced by coaddition of P4. In agreement with our data on mRNA expression (Fig. 1B), treatment with SAHA or TSA alone up-regulated glycodelin protein in a time- and dose-dependent fashion (Fig. 2, C and D, respectively). In all immunoprecipitation experiments, the


FIG. 2. HDACIs and ovarian steroid hormones up-regulate glycodelin protein in a time- and dose-dependent manner. Subconfluent Ishikawa cells were transfected without (MOCK) or with pEGFP-N3 or pcGd1-EGFP using Lipofectamine 2000 (Invitrogen Life Technologies) according to the manufacturer’s instructions and harvested 2 d later for immunoblotting with antiglycodelin or anti-GFP antibody. Bands corresponding to EGFP and Gd1-EGFP are indicated by white and black arrows, respectively (A). Ishikawa cells were cultured without (Ctrl) or with 10 nM E2 alone or 10 nM E2 plus 1 ␮M P4 (E2P4) for the indicated days (B). Ishikawa cells were treated with 2.0 ␮M SAHA or 250 nM TSA for the indicated periods (C) or were cultured in the absence or presence of the indicated concentrations of SAHA or TSA for 3 d (D). Total cell lysates were extracted from the treated cells and subjected to immunoprecipitation and immunoblotting with antiglycodelin (B–D). Each input cell lysate (one to the 10th volume) was subjected to immunoblotting with anti-MAPK antibody (B–D).

input lanes were equally loaded with protein, as determined by immunoblotting with anti-MAPK antibody (Fig. 2, B–D). SAHA and TSA induce LIF mRNA and protein

Besides glycodelin, LIF is known to be up-regulated in response to the rising levels of P4 in luminal and glandular epithelia in the midsecretory phase (16) and therefore is thought to be one of the endometrial epithelial differentiation markers. Similar to glycodelin, the expression level of the LIF gene was significantly increased by treatment with E2P4, SAHA, or TSA for 3 d (Fig. 3A). Consistently, the expression



5369

scriptional activity of several HDACI-susceptible genes through Sp1 sites (17–21). Taking advantage of this information, we constructed a reporter plasmid pGL3 Gd304 in which the promoter (⫺304 to ⫹20 bp) was linked to the luciferase reporter pGL3 (Promega Corp.) and performed luciferase assays. As expected from our present data on the mRNA and protein expression of glycodelin, E2 alone activated the glycodelin promoter, which was enhanced by coaddition of P4 (Fig. 4). The promoter activation by E2P4 was comparable with that by treatment with a low dose of SAHA (125 nm) or TSA (31.3– 62.5 nm) alone. SAHA and TSA alone dramatically stimulated promoter activity in a dose-dependent manner. HDACIs induce morphological changes in Ishikawa cells

FIG. 3. HDACIs and ovarian steroid hormones up-regulate LIF mRNA and protein. Ishikawa cells were cultured without (Ctrl) or with 10 nM E2 plus 1 ␮M P4 (E2P4) or the indicated concentrations of HDACIs for 3 d. Total RNA and total cell lysates derived from the treated cells were subjected to RT-PCR for LIF and GAPDH (A) and immunoprecipitation/immunoblotting with anti-LIF antibody or antiMAPK antibody (B) as described in Materials and Methods. Bars indicate the mean ⫾ SEM relative ratio obtained from three independent experiments (A). Asterisks show significant differences compared with nonstimulated Ishikawa cells (P ⬍ 0.05).

of LIF protein was also up-regulated by the same treatments (Fig. 3B). SAHA and TSA activate the glycodelin promoter

We also focused on the regulatory mechanism of glycodelin gene expression by HDACIs. Gao et al. (12) demonstrated that in the human adenocarcinoma cell line HEC-1B, ligandactivated PR stimulates glycodelin gene expression through functional specificity protein 1 (Sp1) sites present in the glycodelin promoter region between ⫺304 to ⫹20 bp. In addition, HDACIs have been demonstrated to modulate the tran-

FIG. 4. HDACIs activate glycodelin promoter activity. Bars indicate the mean ⫾ SEM relative luciferase activity obtained from three independent experiments. The luciferase activity of pGL3 Gd 304-transfected Ishikawa cells treated with the control vehicle was set at 1.0. Transfection and luciferase assays were performed as described in Materials and Methods. Asterisks indicate significant differences compared with the control (P ⬍ 0.05).

Ishikawa cells became more flattened and widespread when cultured in the presence of E2P4 for 5 d (Fig. 5Ab) compared with the effect of treatment with control vehicles for the same duration (Fig. 5Aa). Moreover, the decrease in the nucleus/cytoplasm ratio was accompanied by an enlargement of the cell size. These characteristic morphological changes induced by treatment with ovarian steroid hormones were also elicited by treatment with SAHA and TSA alone (Fig. 5A, c and d, respectively). HDACIs stimulate glycogen synthesis

Glandular and luminal epithelia dramatically synthesize glycogen in response to the rising levels of P4 after ovulation (22, 23). Intracellular accumulation of glycogen forms subnuclear vacuoles, which is one of characteristic morphological changes observed in the differentiating glandular and luminal epithelia at the P4-dominated early secretory phase (22, 23). We therefore examined whether, in addition to the up-regulation of differentiation markers and morphological changes, HDACIs alone stimulated glycogen synthesis in cultured Ishikawa cells. As shown in Fig. 5B, treatment with E2P4 significantly increased the glycogen content 10 –15% more than control treatment. Likewise, SAHA or TSA alone

5370



FIG. 6. HDACIs and ovarian steroid hormones elicit similar patterns of cell growth in Ishikawa cells. Ishikawa cells were grown in the absence (Control) or the presence of E2 alone, E2P4, 0.25 ␮M SAHA, or 62.5 nM TSA for the indicated culture periods. Each bar represents the mean ⫾ SEM cell number obtained from three independent experiments. Asterisks show significant differences compared with the control Ishikawa cells during the same culture period (P ⬍ 0.05).

Gene silencing of glycodelin by siRNA results in blockade of HDACI-induced differentiation in Ishikawa cells

FIG. 5. HDACIs and ovarian steroid hormones elicit similar morphological changes and stimulate glycogen synthesis in Ishikawa cells. A, Differential interference contrast images of fixed hematoxylin- and eosin-stained Ishikawa cells grown in the absence (a) or the presence of E2P4 (b), 2.0 ␮M SAHA (c), or 250 nM TSA (d) for 3 d. Bar, 20 ␮m. The intracellular glycogen content was measured in Ishikawa cells treated without (Ctrl) or with E2P4 or various concentrations of SAHA and TSA as indicated. Each bar represents the mean ⫾ SEM relative glycogen ratio obtained from three independent experiments (B). Asterisks show significant differences compared with the control (P ⬍ 0.05).

induced glycogen synthesis (Fig. 5B). SAHA exhibited an approximately 40% increment in glycogen synthesis at 2.0 ␮m compared with the control treatment, whereas the stimulating effect of 250 nm TSA was comparable to that of E2P4 (Fig. 5B). HDACIs mimic the effect of E2P4 on cell proliferation in Ishikawa cells at the optimal concentration

It has been reported that E2 has mitogenic activity on Ishikawa cells and that P4 antagonizes the E2 action (24). As shown in Fig. 6, 9 d of cotreatment with P4 could counteract the cell-proliferating effect of E2. Although some attenuation of E2 action by P4 was still observed after 12 d of treatment, prolonged stimulation of E2P4 promoted cell growth compared with control treatment in Ishikawa cells. Notably, SAHA or TSA alone displayed similar proliferation-modulating effects comparable to E2P4 at the optimal concentration (Fig. 6), although high doses of TSA and SAHA induced apoptosis (data not shown).

Lastly, because the production of glycodelin was enhanced by treatment with HDACIs, as presented here, we addressed the question of whether the up-regulated glycodelin per se functionally contributes to HDACI-induced differentiation of Ishikawa cells. As shown in Fig. 7A, glycodelin siRNA, but not GAPDH siRNA, blocked the HDACI-induced expression of glycodelin protein, indicating the validity of the genesilencing effect of the siRNA used in this study. We then examined the effect of glycodelin knockdown on HDACIinduced morphological changes and glycogen synthesis. Ishikawa cells became flattened and widespread (Fig. 7B, b and c), accompanied by an increase in glycogen synthesis (Fig. 7C), when cultured in the presence of E2P4 or SAHA compared with the effect of treatment with control vehicles (Fig. 7, Ba and C). In contrast, treatment with glycodelin siRNA inhibited the SAHA-induced morphological changes and glycogen synthesis (Fig. 7, Bd and C), whereas transfection of GAPDH siRNA provoked no significant alterations in morphological changes (data not shown) and glycogen synthesis (Fig. 7C). Furthermore, glycodelin siRNA, but not GAPDH siRNA, abrogated E2P4- or SAHA-mediated induction of LIF, one of the established glandular epithelial differentiation markers (Fig. 7D). Discussion

We demonstrated in this study that TSA, SAHA, and SBHA, but not HC toxin, could induce the phenotype of a human endometrial adenocarcinoma cell line, Ishikawa, to differentiate to closely resemble normal endometrial epithelium in the absence of ovarian steroid hormones. TSA, SAHA, and SBHA belong to the hydroxamic acid group of HDACIs that possess hydroxamic acid moiety in the metal binding domain, which interacts with the active site of HDACs and thereby inhibits the activity (6). HC toxin is a



5371

cyclic peptide inhibitor that bears a nonhydroxamic acid moiety in the metal binding domain and also possesses a macrocycle containing the surface recognition domain that interacts with residues on the rim of the active site of HDACs (6). These structural differences in the metal binding domain are thought to determine the substrate selectivity targeting individual HDACs; therefore, HDACIs exhibit different potentials to provoke cytodifferentiation (6). This idea, however, seems to be inconsistent with our present data that Scriptaid, which is a hydroxamic acid-type HDACI like TSA and SAHA, could not provoke apparent induction of glycodelin expression. TSA and SAHA have a monocycle in the surface recognition domain, whereas Scriptaid and HC toxin possess a tricycle and a macrocycle in the same domain, respectively (6). Therefore, it is conceivable that such structural discrepancies in the surface recognition domain, instead of the metal binding domain, may also be attributable to the differential inhibition of HDACs, thereby provoking various actions of HDACIs on cytodifferentiation. Glycodelin is a secretory phase dominant glycoprotein that is synthesized by the endometrial epithelium in response to P4 exposure (9). During the periovulatory midcycle, the human endometrium contains no detectable glycodelin (25, 26). However, it appears in endometrial glands 4 –5 d after ovulation, first in some glands, then gradually increasing so that 10 d after ovulation all glands are strongly positive (26). Based on temporal and spatial expressions, glycodelin is widely used as a marker of endometrial epithelial differentiation (9). Glycodelin inhibits egg-sperm binding in a dose-dependent manner (27) and has an immunosuppressive potential to inactivate T cells and natural killer cells (28), suggesting that glycodelin may contribute to contraceptive activity during the latter half of the secretory phase and also may protect the embryonic semiallograft from maternal immune insults. Aberrant expression of glycodelin has been reported in pathological conditions of human endometrium or in endometrium-derived disorders (9). Decreased immunostaining of glycodelin is associated with histologically retarded endometrium, suggesting its clinical relevance to implantation failure (29). Malignant endometrium does not appear to synthesize glycodelin (30), whereas serum glycodelin levels are elevated in patients with advanced endometriosis (31). We have previously reported that TSA alone can induce

FIG. 7. Gene silencing of glycodelin by siRNA results in the blockade of HDACI-induced differentiation in Ishikawa cells. Ishikawa cells were treated with or without 2.0 ␮M SAHA in combination with glycodelin siRNA or GAPDH siRNA for 3 d, and total cell lysates were subjected to immunoprecipitation and immunoblotting with antiglycodelin (A). B, Differential interference contrast images of fixed hematoxylin- and eosin-stained Ishikawa cells grown in the absence

(a) or the presence of E2P4 (b) or in the presence of 2.0 ␮M SAHA without (c) or with (d) glycodelin siRNA. Bar, 20 ␮m. The intracellular glycogen content was measured in Ishikawa cells transfected without (MOCK) or with either GAPDH siRNA or glycodelin siRNA, followed by incubation in the medium without (Ctrl) or with 2 ␮M SAHA (C). Each bar represents the mean ⫾ SEM relative glycogen ratio obtained from three independent experiments. Asterisks show significant differences compared with the corresponding nonstimulated Ishikawa cells (P ⬍ 0.05). Total RNA was extracted from Ishikawa cells treated without (Ctrl) or with either E2P4 or 2 ␮M SAHA in combination with either glycodelin or GAPDH siRNA for 3 d and subjected to RT-PCR analysis for LIF and GAPDH genes (D). A single representative ethidium bromide staining of agarose gel is shown. Bars indicate the mean ⫾ SEM relative ratio of LIF compared with GAPDH, obtained from three independent experimental results. Asterisks show significant differences compared with the nonstimulated cells transfected with GAPDH siRNA (P ⬍ 0.05).

5372


differentiation of human endometrial stromal cells (decidualization) at a relatively high concentration (8). However, our main finding in the previous report is that TSA dramatically accelerates steroid-induced decidualization in a dose-dependent manner (8). In the present study we did not observe such drastic enhancing effects of SAHA and TSA on the steroid-induced differentiation of Ishikawa cells. It is possible that the discrepant effects are due to the cell type or origin, i.e. primary culture vs. cell line, or stroma vs. glandular epithelium. We demonstrated in this study that optimal concentrations of HDACIs affected cell proliferation and differentiation of Ishikawa cells in a similar way as E2P4. The precise mechanism by which HDACIs mimic the action of P4 remains to be elucidated. Because there are no structural similarities between P4 and HDACIs, it is unlikely that HDACIs act as ligands for the PR expressed in Ishikawa cells. One possibility is that HDACIs may augment the PR response to extremely low levels of P4 or P4-like substances present in culture medium, which eventually drives cytodifferentiation. Indeed, several studies have demonstrated that HDACIs can induce or enhance the expression of nuclear steroid receptors (32–35). In addition, it is well known that the phosphorylation of steroid receptors by MAPK increased their activities (5). Because Zhong et al. (36) reported that HDACIs can activate MAPK pathways, HDACIs may positively regulate the activity of the receptors through MAPK activation. However, these possibilities do not seem to be supported by our present findings that neither additive nor synergistic differentiation-inducing effects of TSA and SAHA were observed even upon coaddition of ovarian steroids. We therefore favor a mechanism by which HDACIs do not act through the PR during the differentiation of Ishikawa cells. In the human adenocarcinoma cell line HEC-1B, ligandactivated PR has been demonstrated to stimulate glycodelin gene expression through functional Sp1 sites present in the glycodelin promoter region between ⫺304 and ⫹20 bp (12). TSA and SAHA are known to activate the promoter of several genes, including p21/wild-type p53-activated fragment 1/cyclin-dependent kinase-interacting protein 1, human telomerase reverse transcriptase, mitochondrial 3-hydroxy3-methyl-glutaryl-coenzyme A synthase, and inhibitor of cyclin-dependent kinase 4d through the Sp1 site(s) in some types of cells that do not express PRs (17–21). In agreement, we here showed that in the absence of P4, TSA and SAHA could activate the glycodelin promoter harboring functional Sp1 sites. In addition to glycodelin, we found that TSA and SAHA induced LIF expression in Ishikawa cells. In the 666-bp LIF promoter region (37), a single putative Sp1 site is predicted by the sequence motif search program, TFSEARCH, version 1.3 (www.cbrc.jp/research/db/TFSEARCHJ.html). It is therefore conceivable that HDACIs stimulate LIF gene expression through the single Sp1 site. HDACIs modify the local chromatin structure and thereby primarily regulate the promoter of a set of genes to initiate differentiation, whose gene products, in turn, may affect the expression of downstream genes and finally accomplish differentiation (7). We found that HDACIs directly activated the minimal glycodelin promoter and thereby induced mRNA and protein expression of glycodelin. Our knockdown experiments using siRNA demonstrated that down-regulation


of glycodelin abrogated the differentiation-inducing effects of HDACI on Ishikawa cells, suggesting that glycodelin may behave as an effecter molecule to drive endometrial epithelial differentiation. In addition, reduced expression of glycodelin resulted in the impairment of E2P4-induced LIF expression. These results collectively indicate that glycodelin is the critical determinant of endometrial epithelial differentiation. In agreement, it has been suggested that glycodelin plays a role as a differentiation-related glandular morphogen (9). Glycodelin is exclusively present in glandular structures of many tissues, including endometrium, lobular and ductal epithelium of the breast, eccrine sweat glands, and parabronchial glands (38, 39). Furthermore, overexpression of glycodelin in glycodelin-negative breast cancer cells results in the formation of gland-like structures, restricted proliferation, and induction of other mammary epithelial markers (39). We propose a model in which HDACIs up-regulate the glycodelin gene, presumably through the Sp1 site(s), without the assistance of ligand-activated PR, and HDACI-induced glycodelin per se, in turn, inhibits cell growth and drives endometrial epithelial differentiation. In this context, HDACIs potentiate cytodifferentiation via induced glycodelin in Ishikawa cells. The present study may provide clues for possible differentiation therapies using HDACIs or in combination with P4 to target histone acetylation/deacetylation and glycodelin for endometrium-derived diseases such as endometrial cancer and endometriosis. Indeed, phase I and II clinical trials have recently demonstrated that SAHA is a promising anticancer drug (40 – 42). Our results also implicate the use of HDACIs in investigation of molecular mechanisms underlying steroid-induced endometrial epithelial differentiation. Acknowledgments We are grateful to Ms. Rei Sakurai for technical assistance, and to Ms. Shino Kuwabara for secretarial work. We also thank Drs. Masato Nishida (National Kasumigaura Hospital, Ibaragi, Japan) for Ishikawa cells, and Toshihide Tajima (Keio University, Tokyo, Japan) for glycogen analysis. Received March 28, 2005. Accepted August 18, 2005. Address all correspondence and requests for reprints to: Dr. Tetsuo Maruyama, Department of Obstetrics and Gynecology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku, Tokyo 160-8582, Japan. E-mail: [email protected]. This work was supported in part by Grants-in-Aid for Scientific Research C16591683 (to H.U.) and C13671743 (to T.M.), Grants-in-Aid for Exploratory Research 15659399 (to T.M.) from the Japan Society for the Promotion of Science, grants from Keio University for the Encouragement of Young Medical Scientists, Keio Medical Association (to H.U.), and Mitsukoshi Medical Foundation (to T.M.).

References 1. Punyadeera C, Verbost P, Groothuis P 2003 Oestrogen and progestin responses in human endometrium. J Steroid Biochem Mol Biol 84:393– 410 2. de Ziegler D, Fanchin R 2000 Progesterone and progestins: applications in gynecology. Steroids 65:671– 679 3. Ramirez PT, Frumovitz M, Bodurka DC, Sun CC, Levenback C 2004 Hormonal therapy for the management of grade 1 endometrial adenocarcinoma: a literature review. Gynecol Oncol 95:133–138 4. McKenna NJ, O’Malley BW 2002 Combinatorial control of gene expression by nuclear receptors and coregulators. Cell 108:465– 474 5. Lange CA 2004 Making sense of cross-talk between steroid hormone receptors


6. 7. 8.

9. 10. 11. 12. 13.

14. 15.

16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

and intracellular signaling pathways: who will have the last word? Mol Endocrinol 18:269 –278 Miller TA, Witter DJ, Belvedere S 2003 Histone deacetylase inhibitors. J Med Chem 46:5097–5116 Marks PA, Richon VM, Miller T, Kelly WK 2004 Histone deacetylase inhibitors. Adv Cancer Res 91:137–168 Sakai N, Maruyama T, Sakurai R, Masuda H, Yamamoto Y, Shimizu A, Kishi I, Asada H, Yamagoe S, Yoshimura Y 2003 Involvement of histone acetylation in ovarian steroid-induced decidualization of human endometrial stromal cells. J Biol Chem 278:16675–16682 Seppa¨la¨ M, Taylor RN, Koistinen H, Koistinen R, Milgrom E 2002 Glycodelin: a major lipocalin protein of the reproductive axis with diverse actions in cell recognition and differentiation. Endocr Rev 23:401– 430 Takai N, Desmond JC, Kumagai T, Gui D, Said JW, Whittaker S, Miyakawa I, Koeffler HP 2004 Histone deacetylase inhibitors have a profound antigrowth activity in endometrial cancer cells. Clin Cancer Res 10:1141–1149 Nishida M 2002 The Ishikawa cells from birth to the present. Hum Cell 15:104 –117 Gao J, Mazella J, Seppa¨la¨ M, Tseng L 2001 Ligand activated hPR modulates the glycodelin promoter activity through the Sp1 sites in human endometrial adenocarcinoma cells. Mol Cell Endocrinol 176:97–102 Kao LC, Germeyer A, Tulac S, Lobo S, Yang JP, Taylor RN, Osteen K, Lessey BA, Giudice LC2003 Expression profiling of endometrium from women with endometriosis reveals candidate genes for disease-based implantation failure and infertility. Endocrinology 144:2871–2880 Chen HF, Chao K-H, Shew J-Y, Yang Y-S, Ho H-N 2004 Expression of leukemia inhibitory factor and its receptor is not altered in the decidua and chorionic villi of human anembryonic pregnancy. Hum Reprod 19:1647–1654 Arkwright PD, Rademacher TW, Dwek RA, Redman CW 1993 Pre-eclampsia is associated with an increase in trophoblast glycogen content and glycogen synthase activity, similar to that found in hydatidiform moles. J Clin Invest 91:2744 –2753 Vogiagis D, Salamonsen LA 1999 Review: The role of leukaemia inhibitory factor in the establishment of pregnancy. J Endocrinol 160:181–190 Sowa Y, Orita T, Hiranabe-Minamikawa S, Nakano K, Mizuno T, Nomura H, Sakai T 1999 Histone deacetylase inhibitor activates the p21/WAF1/Cip1 gene promoter through the Sp1 sites. Ann NY Acad Sci 886:195–199 Huang L, Sowa Y, Sakai T, Pardee AB 2000 Activation of the p21WAF1/CIP1 promoter independent of p53 by the histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA) through the Sp1 sites. Oncogene 19:5712–5719 Takakura M, Kyo S, Sowa Y, Wang Z, Yatabe N, Maida Y, Tanaka M, Inoue M 2001 Telomerase activation by histone deacetylase inhibitor in normal cells. Nucleic Acids Res 29:3006 –3011 Camarero N, Nadal A, Barrero MJ, Haro D, Marrero PF 2003 Histone deacetylase inhibitors stimulate mitochondrial HMG-CoA synthase gene expression via a promoter proximal Sp1 site. Nucleic Acids Res 31:1693–1703 Yokota T, Matsuzaki Y, Miyazawa K, Zindy F, Roussel MF, Sakai T 2004 Histone deacetylase inhibitors activate INK4d gene through Sp1 site in its promoter. Oncogene 23:5340 –5349 Kohorn EI, Rice SI, Hemperly S, Gordon M 1972 The relation of the structure of progestational steroids to nucleolar differentiation in human endometrium. J Clin Endocrinol 34:257–264 Spornitz UM 1992 The functional morphology of the human endometrium and decidua. Adv Anat Embryol Cell Biol 124:1–99 Croxtall JD, Elder MG, White JO 1990 Hormonal control of proliferation in the Ishikawa endometrial adenocarcinoma cell line. J Steroid Biochem 35:665– 669 Waites GT, Wood PL, Walker RA, Bell SC 1988 Immunohistological localization of human endometrial secretory protein, ‘pregnancy-associated endo-


26.

27. 28. 29. 30.

31.

32. 33.

34.

35. 36. 37. 38. 39. 40.

41.

42.

5373

metrial ␣2-globulin’ (␣2-PEG), during the menstrual cycle. J Reprod Fertil 82:665– 672 Seppa¨la¨ M, Wahlstrom T, Julkunen M, Vartiainen E, Huhtala M-L 1988 Endometrial proteins as indicators of endometrial function. In: Tomoda Y, Mizutani S, Narita O, Klopper A, eds. Placental and endometrial proteins: basic and clinical aspects. Utrecht: VNU Science Press; 35– 42 Oehninger S, Coddington CC, Hodgen GD, Seppa¨la¨ M 1995 Factors affecting fertilization: endometrial placental protein 14 reduces the capacity of human spermatozoa to bind to the human zona pellucida. Fertil Steril 63:377–383 Okamoto N, Uchida A, Takakura K, Kariya Y, Kanzaki H, Riittinen L, Koistinen R, Seppa¨la¨ M, Mori T 1991 Suppression by human placental protein 14 of natural killer cell activity. Am J Reprod Immunol 26:137–142 Klentzeris LD, Bulmer JN, Seppa¨la¨ M, Li TC, Warren MA, Cooke ID 1994 Placental protein 14 in cycles with normal and retarded endometrial differentiation. Hum Reprod 9:394 –398 Wood PL, Waites GT, MacVicar J, Davidson AC, Walker RA, Bell SC 1988 Immunohistological localization of pregnancy-associated endometrial ␣2globulin (␣2-PEG) in endometrial adenocarcinoma and effect of medroxyprogesterone acetate. Br J Obstet Gynaecol 95:1292–1298 Telimaa S, Kauppila A, Ronnberg L, Suikkari AM, Seppa¨la¨ M 1989 Elevated serum levels of endometrial secretory protein PP14 in patients with advanced endometriosis. Suppression by treatment with danazol and high-dose medroxyprogesterone acetate. Am J Obstet Gynecol 161:866 – 871 Ruh MF, Tian S, Cox LK, Ruh TS 1999 The effects of histone acetylation on estrogen responsiveness in MCF-7 cells. Endocrine 11:157–164 Yang X, Phillips DL, Ferguson AT, Nelson WG, Herman JG, Davidson NE 2001 Synergistic activation of functional estrogen receptor (ER)-␣ by DNA methyltransferase and histone deacetylase inhibition in human ER-␣-negative breast cancer cells. Cancer Res 61:7025–7029 Keen JC, Yan L, Mack KM, Pettit C, Smith D, Sharma D, Davidson NE 2003 A novel histone deacetylase inhibitor, Scriptaid, enhances expression of functional estrogen receptor ␣ (ER) in ER negative human breast cancer cells in combination with 5-aza 2⬘-deoxycytidine. Breast Cancer Res Treat 81:177–186 Jang ER, Lim SJ, Lee ES, Jeong G, Kim TY, Bang YJ, Lee JS 2004 The histone deacetylase inhibitor trichostatin A sensitizes estrogen receptor ␣-negative breast cancer cells to tamoxifen. Oncogene 23:1724 –1736 Zhong S, Goto H, Inagaki M, Dong Z 2003 Phosphorylation at serine 28 and acetylation at lysine 9 of histone H3 induced by trichostatin A. Oncogene 22:5291–5297 Bamberger AM, Erdmann I, Bamberger CM, Jenatschke SS, Schulte HM 1997 Transcriptional regulation of the human ‘leukemia inhibitory factor’ gene: modulation by glucocorticoids and estradiol. Mol Cell Endocrinol 127:71–79 Kamarainen M, Halttunen M, Koistinen R, von Boguslawsky K, von Smitten K, Andersson LC, Seppa¨la¨ M 1999 Expression of glycodelin in human breast and breast cancer. Int J Cancer 83:738 –742 Kamarainen M, Seppa¨la¨ M, Virtanen I, Andersson LC 1997 Expression of glycodelin in MCF-7 breast cancer cells induces differentiation into organized acinar epithelium. Lab Invest 77:565–573 Kelly WK, Richon VM, O’Connor O, Curley T, MacGregor-Curtelli B, Tong W, Klang M, Schwartz L, Richardson S, Rosa E, Drobnjak M, Cordon-Cordo C, Chiao JH, Rifkind R, Marks PA, Scher H 2003 Phase I clinical trial of histone deacetylase inhibitor: suberoylanilide hydroxamic acid administered intravenously. Clin Cancer Res 9:3578 –3588 Blumenschein G, Lu C, Kies M, Glisson B, Papadimitrakopoulou V, Zinner R, Kim E, Gillenwater AJCWH 2004 Phase II clinical trial of suberoylanilide hydroxamic acid (SAHA) in patients (pts) with recurrent and/or metastatic head and neck cancer (SCCHN). J Clin Oncol 22:5578 Johnstone RW 2004 Suberanilohydroxamic acid. Aton Pharma IDrugs 7:674 – 682

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.