Human ovarian carcinoma cells: Histone deacetylase inhibitors exhibit antiproliferative activity and potently induce apoptosis

Authors

  • Noriyuki Takai M.D., Ph.D.,

    1. Division of Hematology/Oncology, Cedars-Sinai Medical Center/University of California–Los Angeles School of Medicine, Los Angeles, California
    2. Department of Obstetrics and Gynecology, Faculty of Medicine, Oita University, Oita, Japan
    Search for more papers by this author
    • The first two authors contributed equally to this article.

  • Norihiko Kawamata M.D.,

    1. Division of Hematology/Oncology, Cedars-Sinai Medical Center/University of California–Los Angeles School of Medicine, Los Angeles, California
    Search for more papers by this author
    • The first two authors contributed equally to this article.

  • Dorina Gui M.D.,

    1. Department of Pathology, University of California–Los Angeles School of Medicine Center for Health Sciences, Los Angeles, California
    Search for more papers by this author
  • Jonathan W. Said M.D.,

    1. Department of Pathology, University of California–Los Angeles School of Medicine Center for Health Sciences, Los Angeles, California
    Search for more papers by this author
  • Isao Miyakawa M.D., Ph.D.,

    1. Department of Obstetrics and Gynecology, Faculty of Medicine, Oita University, Oita, Japan
    Search for more papers by this author
  • H. Phillip Koeffler M.D.

    Corresponding author
    1. Division of Hematology/Oncology, Cedars-Sinai Medical Center/University of California–Los Angeles School of Medicine, Los Angeles, California
    • Division of Hematology/Oncology, Cedars-Sinai Medical Center/University of California–Los Angeles School of Medicine, 8700 Beverly Boulevard, Los Angeles, CA 90048
    Search for more papers by this author
    • Fax: (310) 423-0443

    • H. Phillip Koeffler holds the Mark Goodson Endowed Chair of Oncology Research and is a member of the Jonsson Cancer Center and the Molecular Biology Institute at the University of California–Los Angeles (Los Angeles, CA).


Abstract

BACKGROUND

Histone deacetylase inhibitors (HDACIs) can inhibit proliferation, stimulate apoptosis, and induce cell cycle arrest in malignant cells.

METHODS

The authors investigated the effects of four HDACIs on nine ovarian carcinoma cell lines in vitro and in vivo. Ovarian carcinoma cells were treated with a variety of HDACIs, and their effects on cell growth, the cell cycle, apoptosis, and related events were investigated. The ability of valproic acid (VPA) to inhibit the growth of ovarian tumors in immunodeficient mice was also assessed.

RESULTS

Clonogenic assays showed that all ovarian carcinoma cell lines were sensitive to the growth-inhibitory effects of the HDACIs. Cell cycle analysis indicated that their exposure to HDACIs decreased the proportion of cells in S phase and increased the proportion of cells in the G0/G1 and/or G2/M phases of the cell cycle. Terminal deoxynucleotidyltransferase-mediated uridine triphosphate end-labeling assays demonstrated that HDACIs induced apoptosis, which occurred in concert with alterations in the expression of genes related to apoptosis, cell growth, and malignant phenotype, including the activation of caspase-9 and caspase-3. Chromatin immunoprecipitation analysis revealed a notable increase in levels of acetylated histones associated with the p21 promoter after treatment with suberoylanilide bishydroxamine. In addition, in experiments involving nude mice, VPA significantly inhibited human ovarian tumor growth without toxic side effects.

CONCLUSIONS

The results of the current study suggest that HDACIs may be particularly effective in the treatment of ovarian tumors. Cancer 2004. © 2004 American Cancer Society.

One of the most important mechanisms in chromatin remodeling is the posttranslational modification of the N-terminal tails of histones by acetylation, which contributes to a ‘histone code’ that determines the activity of target genes.1 Transcriptionally silent chromatin is composed of nucleosomes in which the histones have low levels of acetylation at the lysine residues of their amino-terminal tails. Acetylation of histone proteins neutralizes the positive charge on lysine residues and disrupts nucleosome structure, allowing unfolding of the associated DNA and subsequent access by transcription factors, resulting in changes in gene expression.

Acetylation of core nucleosomal histones is regulated by the opposing activities of histone acetyltransferases and deacetylases (HDACs). HDACs catalyze the removal of acetyl groups on the amino-terminal lysine residues of core nucleosomal histones, and this activity is generally associated with transcriptional repression. Aberrant recruitment of HDAC activity has been associated with the development of certain human malignancies.2 Transcription factors such as Mad-1, bcl-6, and ETO also have been shown to assemble HDAC-dependent transcriptional repressor complexes.3–5

HDAC inhibitors (HDACIs), such as trichostatin A (TSA) and sodium butyrate (NaB), can inhibit malignant cell growth in vitro6 and in vivo,7 bring about the reversion of oncogene-transformed cell morphology,8 induce apoptosis,9 and enhance cell differentiation.10 Several classes of HDACIs have been identified,11 including short-chain fatty acids (e.g., butyrates and valproic acid [VPA]), organic hydroxamic acids (e.g., TSA and suberoyl anilide bishydroxamine [SAHA]), cyclic tetrapeptides (e.g., trapoxin), and benzamides (e.g., MS-275).11 The structure of SAHA is related to that of TSA,11 a natural product isolated from Streptomyces hygroscopicus that was used initially as an antifungal antibiotic. Phenylbutyrate has been used as a single agent in the treatment of β-thalassemia, toxoplasmosis, and malaria.

SAHA represents a novel therapeutic approach to the treatment of malignant disease and is in Phase I of clinical trials for the treatment of a variety of solid and hematologic tumors. VPA is an established agent for use in the long-term treatment of epilepsy. Some HDACIs (e.g., TSA and trapoxin) are of limited therapeutic use, due to poor bioavailability in vivo and toxic side effects at high doses. NaB and phenylbutyrate are degraded rapidly after intravenous administration, and as a result, doses > 400 mg/kg per day are required for these agents to be effective.12 Furthermore, these compounds are not specifically active against HDACs, as they also inhibit the phosphorylation and methylation of other proteins, as well as DNA methylation.13

The current study was designed to define the biologic and therapeutic effects of HDACIs in the treatment of ovarian carcinoma. We focused particularly on SAHA and VPA, which are recognized as the least toxic HDACIs. We examined whether these compounds were able to mediate cell growth inhibition, cell cycle arrest, apoptosis, and the expression of genes related to malignant phenotypes in a variety of ovarian carcinoma cell lines. Furthermore, we examined whether SAHA was able to induce the accumulation of acetylated histones in the chromatin of the p21WAF1 gene in human ovarian carcinoma cells. Finally, we tested the ability of VPA to inhibit the growth of the ovarian carcinoma cell line SK-OV-3 in vivo.

MATERIALS AND METHODS

Cell Lines

The human cell lines SK-OV-3, OVCAR-3, TOV-21G, OV-90, and TOV-112D were obtained from the American Type Culture Collection (Manassas, VA). OVCA420, OVCA429, OVCA432, and OVCA433 were kindly provided by Dr. Robert C. Bast, Jr. (The University of Texas M. D. Anderson Cancer Center, Houston, TX). Cells were maintained as monolayers at 37 °C (5% CO2/air atmosphere) in RPMI-1640 (Gibco, Rockville, MD) containing 10% heat-inactivated fetal bovine serum (Omega, Tarzana, CA).

Histone Deacetylase Inhibitors

SAHA was generously provided by Dr. Victoria Richon (Aton Pharma, Tarrytown, NY), and VPA, TSA, and NaB were obtained from Sigma (St. Louis, MO). SAHA was dissolved in dimethylsulfoxide (DMSO; Sigma) at a concentration of 5 × 10−2 M and then diluted in phosphate-buffered saline (PBS) to a concentration of 5 × 10− 3 M before use. VPA, TSA, and NaB were dissolved in PBS at concentrations of 5 M, 5 × 10−3 M, and 5 M, respectively. Diluent alone (DMSO or PBS) was added to culture media as a negative control.

Clonogenic Assay in Soft Agar

The effect of HDACIs on the clonogenic growth of ovarian carcinoma cells was determined by dose-response studies in soft agar as previously described.14 All assays were performed independently at least three times per experimental point in triplicate dishes.

Cell Cycle Analysis by Flow Cytometry

Flow cytometric analysis of the cell cycle was performed after 3 days of culturing either with or without HDACIs, as is described elsewhere.14 The analysis, performed immediately after staining, was carried out using the CELLFit program (Becton Dickinson, Franklin Lakes, NJ) which calculated the S-phase fraction using an RFit model.

Measurement of Apoptosis

DNA strand breaks were identified via terminal deoxynucleotidyltransferase-mediated uridine triphosphate end-labeling (TUNEL) analysis, using the In Situ Cell Death Detection Kit in accordance with the manufacturer's instructions (Boehringer-Mannheim, Mannheim, Germany).

Western Blotting

Expression of specific proteins was detected by Western blot analysis. Whole lysates (40 μg) were resolved on a 4–15% sodium dodecyl sulfate–polyacrylamide gel and transferred to an Immobilon polyvinylidene difuride membrane (Amersham, Arlington Heights, IL). Antiacetylated H3 polyclonal antibodies (1:1000 dilution; Upstate Biotechnology, Lake Placid, NY), antiacetylated H4 polyclonal antibodies (1:1000 dilution; Upstate Biotechnology), anti-p21WAF1 monoclonal antibody (MoAb) (Ab-1, 1:1000 dilution; Oncogene, San Diego, CA), anti-p27KIP1 polyclonal antibodies (C-19, 1:1000 dilution; Santa Cruz Biotechnology, Santa Cruz, CA), anti-E-cadherin MoAb (G-10, 1:1000 dilution; Santa Cruz Biotechnology), anti-Bcl-2 MoAb (100, 1:1000 dilution; Santa Cruz Biotechnology), anti-Bax polyclonal antibodies (N20, 1:1000 dilution; Santa Cruz Biotechnology), cyclin D1 MoAb (A-12, 1:1000 dilution; Santa Cruz Biotechnology), cyclin D2 polyclonal antibodies (M-20, 1:1000 dilution; Santa Cruz Biotechnology), anti-caspase-3 antibodies (Pharmingen, San Diego, CA), anti-caspase-9 antibodies (Santa Cruz Biotechnology), anti-poly adenosine diphosphate-ribose polymerase (PARP) antibodies (Santa Cruz Biotechnology), and anti-glyceraldehyde-3-phosphate dehydrogenase MoAb (Research Diagnostics, Flanders, NJ) were used. Blots were developed using an enhanced chemiluminescence (ECL) kit (Amersham).

Chromatin Immunoprecipitation Assay

Cells were plated at a density of 1 × 106 cells per 10 cm dish and incubated overnight at 37 °C with 5% CO2. After 12 hours, cells were cultured in either 5 ×10−6 M SAHA or control diluent for 72 hours. Immunoprecipitated DNA (both immunoprecipitation samples and Input) was prepared using Acetyl-Histone H3/H4 Chromatin Immunoprecipitation (ChIP) Assay Kits (Upstate Biotechnology, Lake Placid, NY) and analyzed via polymerase chain reaction (PCR). β-actin or p21WAF1-specific primers were used to amplify DNA isolated from ChIP experiments and Input samples. The optimal reaction conditions for PCR were determined for each primer pair. Primers were denatured at 95 °C for 1 minute, annealed at 66 °C for 1 minute, and elongated at 72 °C for 1 minute. PCR products were analyzed by 2.5% agarose/ethidium bromide gel electrophoresis. The primer pair used for p21WAF1 ChIP analysis included 5′-GGT GTC TAG GTG CTC CAG GT-3′ (forward) and 5′-GCA CTC TCC AGG AGG ACA CA-3′ (reverse). The primers used for β-actin ChIP analysis were 5′-GCC AGC TCT CGC ACT CTG TT-3′ (forward) and 5′-AGA TCG CAA CCG CCT GGA AC-3′ (reverse).

In Vivo Animal Treatment Protocol

Ten 6-week-old immunodeficient beige/nude/xid nu/nu female mice were purchased from Harlan Sprague Dawley, Inc. (Indianapolis, IN), and maintained under pathogen-free conditions with irradiated chow. In one experiment, 5 ×106 SK-OV-3 cells in 0.1 mL Matrigel (Collaborative Biomedical Products, Bedford, MA) was bilaterally and subcutaneously injected into the trunk of 10 mice, leading to the formation of 2 tumors per animal. Treatment was started on the day after the injection of these human ovarian carcinoma cells and was discontinued after 5 weeks. Cohorts (5 mice per group) received either diluent only (control group) or VPA (10 mg per day) intraperitoneally for 5 days per week. Tumors were measured every week with vernier calipers. Tumor volume was calculated using the following formula: volume = length × width × height × 0.5236. At the end of the experiments (after 5 weeks), blood specimens were collected from the orbital sinus for serum chemistry assays and blood studies, which were performed using the Dupont Analyst Benchtop Chemistry System (Dade International, Newark, DE) and the Serono-Baker 9000 Diff (Biochem Immuno-Systems, Allentown, PA), respectively. Animals were sacrificed by CO2 asphyxiation, after which careful resection was performed and tumor weights were measured. Tumor, bone marrow, liver, lung, spleen, and kidney specimens were fixed and stained for histologic analysis. All animal experiments were performed in compliance with National Institutes of Health guidelines.

Histology

Tumor specimens and specimens from normal organs were fixed in 10% neutral buffered formalin and embedded in paraffin before histologic sectioning. Sections were stained with hematoxylin and eosin, and tumor necrosis and fibrosis were evaluated. Specimens from normal organs were evaluated for evidence of toxic damage. Control samples included tumor and organ specimens from mice not subjected to treatment.

Immunohistochemistry

Immunohistochemistry studies were performed on formalin-fixed sections of tissue specimens. Sections were pretreated with trypsin (10 mg per 50 mL in Tris buffer, pH 8.1) for 10 minutes at 37 °C, followed by anti-p21WAF1 MoAb (1:1000 dilution in PBS; Oncogene) for 30 minutes. Slides then were washed in PBS and incubated sequentially for 15 minutes with peroxidase-conjugated swine anti-mouse immunoglobulin G (1:50 dilution; Dako, Carpinteria, CA). Staining was performed using a Dako autostainer. Localization of reaction products was performed using the diaminobenzidene reaction.

Statistical Analysis

All numeric data were expressed as average values ± standard deviations. The statistical significance of differences among tumors in mice was analyzed via the nonparametric Mann–Whitney U test, which was performed using STAT VIEW software (Abacus Concept, Berkeley, CA). For all other experiments, significance was determined by conducting a paired Student t test.

RESULTS

Effect of Histone Deacetylase Inhibitors on the Clonal Proliferation of Ovarian Carcinoma Cells In Vitro

To study the effect of the HDACIs (SAHA, VPA, TSA, and NaB) on the clonogenic growth of ovarian carcinoma cell lines, a two-layer soft agar system was used. SAHA inhibited clonal proliferation of the SK-OV-3 ovarian carcinoma cell line in a dose-dependent manner (Fig. 1). Similar results were observed in the other eight ovarian carcinoma cell lines studied (data not shown). Dose-response clonogenic studies were performed for all nine ovarian carcinoma cell lines, which were tested against SAHA, VPA, TSA, and NaB. The effective HDACI dose that inhibited 50% clonal growth (ED50) of the ovarian carcinoma cell lines was calculated (Table 1). ED50 doses ranged between 8.1 × 10−7 and 3.9 × 10−6 M for SAHA, between 4.7 × 10−4 and 2.4 × 10−3 M for VPA, between 7.4 × 10−8 and 2.7 × 10−7 M for TSA, and between 6.3 × 10−4 and 7.1 × 10−3 M for NaB. SK-OV-3 cells were the most sensitive to the inhibitory effects of HDACIs (Table 1).

Figure 1.

Dose-response effects of suberoylanilide bishydroxamine (SAHA), a histone deacetylase inhibitor, on clonal proliferation of SK-OV-3 human ovarian carcinoma cells. Dose-response clonogenic assay of human ovarian carcinoma cell line SK-OV-3 cultured with various concentrations of SAHA. Data points represent mean values ± standard deviations from three independent experiments performed in triplicate dishes.

Table 1. HDACIs: Inhibition of Clonal Proliferation (ED50) of Ovarian Carcinoma Cell Linesa
Cell lineED50
SAHAVPATSANaB
  • HDACI: histone deacetylase inhibitors; SAHA: suberoylanilide bishydroxamine; VPA: valproic acid; TSA: trichostatin A; NaB: sodium butyrate.

  • a

    Numbers represent HDACI concentrations that induced 50% inhibition of clonal growth (ED50) in the ovarian carcinoma cell lines. The ED50 was determined by plotting the inhibition of colony formation of ovarian carcinoma cells against increasing concentrations of each HDACI (SAHA: 1 × 10−7 to 5 × 10−6 M [Fig. 1]); VPA: 1 × 10−4 to 5 × 10−3 M; TSA: 1 × 10−8 to 3 × 10−7 M; NaB; 1 × 10−4 to 5 × 10−3 M).

SK-OV-38.1 × 10−74.7 × 10−47.4 × 10−86.3 × 10−4
OVCAR-38.2 × 10−77.8 × 10−48.1 × 10−88.1 × 10−4
TOV-21G3.9 × 10−69.3 × 10−42.7 × 10−77.1 × 10−3
OV-903.5 × 10−62.4 × 10−31.9 × 10−74.6 × 10−3
TOV-112D9.2 × 10−74.7 × 10−49.8 × 10−83.9 × 10−3
OVCA4203.3 × 10−66.6 × 10−41.7 × 10−72.1 × 10−3
OVCA4291.2 × 10−67.5 × 10−41.1 × 10−71.3 × 10−3
OVCA4322.8 × 10−68.3 × 10−49.5 × 10−81.9 × 10−3
OVCA4339.5 × 10−78.2 × 10−48.8 × 10−84.0 × 10−3

Cell Cycle Analysis of Ovarian Carcinoma Cells after Exposure to Histone Deacetylase Inhibitors

The effect of each of the HDACIs on the cell cycle of the ovarian carcinoma cells was assessed. Ovarian carcinoma cells cultured for 3 days in the presence of HDACIs showed an accumulation in the G1/G0 and/or G2/M phases of the cell cycle, with a concomitant decrease in the proportion of cells in S phase (Fig. 2, Table 2). For example, 53% of all untreated SK-OV-3 cells were in G0/G1, compared with 20% of all cells cultured in the presence of SAHA (5 × 10−6 M) and 80% of all cells cultured in the presence of VPA (5 ×10−3 M). Seven percent of all SK-OV-3 untreated cells were in G2/M, compared with 55% of cells treated with SAHA (5 × 10−6 M) and 5% of cells treated with VPA (5 ×10−3 M). These results were representative of all HDACIs tested.

Figure 2.

Cell cycle analysis of SK-OV-3 cells by flow cytometry. SK-OV-3 cells were cultured with suberoylanilide bishydroxamine (SAHA; 5 × 10−6 M), valproic acid (VPA; 5 × 10−3 M), trichostatin A (TSA; 3 × 10−7 M), or sodium butyrate (NaB; 5 × 10−3 M) for 72 hours, harvested, and stained with propidium iodide. Control cells were treated with vehicle alone. Cell cycle analysis was performed using flow cytometry (see Materials and Methods). Data points represent mean values ± standard deviations from three independent experiments. *: P < 0.05 (as determined by the Student t test) for the difference between the experimental group and the control group.

Table 2. Cell Cycle Changes Mediated by HDACIs in Ovarian Carcinoma Cell Linesa
Cell lineControlSAHA (5 × 10−6 M)VPA (5 × 10−3 M)TSA (3 × 10−7 M)NaB (5 × 10−3 M)
  • HDACI: histone deacetylase inhibitors; SAHA: suberoylanilide bishydroxamine; VPA: valproic acid; TSA: trichostatin A; NaB: sodium butyrate.

  • a

    Ovarian carcinoma cells were plated in triplicate wells and grown either in the presence or absence (control) of HDACIs for 3 days at the indicated concentration, and the cell cycle distribution was subsequently measured. Numbers indicate mean percentages ± standard deviations.

SK-OV-3     
 G0/G153 ± 820 ± 780 ± 1632 ± 963 ± 21
 S40 ± 325 ± 1015 ± 519 ± 816 ± 5
 G2/M7 ± 155 ± 115 ± 449 ± 2021 ± 9
OVCAR-3     
 G0/G144 ± 1227 ± 659 ± 1930 ± 1755 ± 31
 S42 ± 831 ± 923 ± 718 ± 412 ± 11
 G2/M14 ± 342 ± 1118 ± 852 ± 2133 ± 10
TOV-21G     
 G0/G147 ± 1036 ± 1061 ± 2238 ± 1557 ± 26
 S38 ± 711 ± 520 ± 523 ± 1113 ± 7
 G2/M15 ± 153 ± 2119 ± 939 ± 2130 ± 15
OV-90     
 G0/G150 ± 1224 ± 969 ± 3029 ± 1060 ± 31
 S38 ± 1120 ± 920 ± 920 ± 718 ± 9
 G2/M12 ± 356 ± 1711 ± 551 ± 1922 ± 11
TOV-112D     
 G0/G142 ± 828 ± 970 ± 2525 ± 758 ± 20
 S45 ± 723 ± 519 ± 1128 ± 616 ± 11
 G2/M13 ± 249 ± 1211 ± 747 ± 926 ± 10
OVCA420     
 G0/G147 ± 1539 ± 1073 ± 3831 ± 1661 ± 18
 S44 ± 1022 ± 417 ± 1030 ± 2117 ± 9
 G2/M9 ± 539 ± 1110 ± 539 ± 2022 ± 3
OVCA429     
 G0/G151 ± 928 ± 559 ± 1839 ± 1059 ± 10
 S42 ± 829 ± 728 ± 719 ± 1123 ± 9
 G2/M7 ± 243 ± 1813 ± 942 ± 1718 ± 5
OVCA432     
 G0/G143 ± 1135 ± 1566 ± 2631 ± 860 ± 24
 S43 ± 815 ± 619 ± 730 ± 1518 ± 9
 G2/M14 ± 450 ± 2215 ± 1039 ± 1622 ± 11
OVCA433     
 G0/G144 ± 927 ± 1874 ± 1725 ± 758 ± 18
 S41 ± 926 ± 1517 ± 930 ± 1521 ± 11
 G2/M15 ± 647 ± 229 ± 345 ± 2021 ± 7

Effect of Histone Deacetylase Inhibitors on the Induction of Apoptosis

The strong antiproliferative effect of HDACIs on ovarian carcinoma cells observed in vitro may be caused, in part, by the induction of apoptosis. To test this hypothesis, we performed the TUNEL assay on ovarian carcinoma cell lines treated with HDACIs for 3 days. SAHA induced apoptosis in a dose-dependent manner, with 56% of all SK-OV-3 cells undergoing apoptosis after 3 days of culturing in the presence of 5 ×10−6 M SAHA (Fig. 3). Exposure to SAHA (5 ×10−6 M) for 3 days induced apoptosis in each of the 9 ovarian carcinoma cell lines studied (SK-OV-3, 56%; OVCA433, 49%; OVCA432, 46%; TOV-112D, 45%; OVCA420, 42%; TOV-21G, 40%; OVCAR-3, 39%; OVCA429, 39%; and OV-90, 34%). These findings were representative of all HDACIs tested (Table 3).

Figure 3.

Induction of apoptosis by histone deacetylase inhibitors. SK-OV-3 cells were treated for 3 days with various concentrations of suberoylanilide bishydroxamine (SAHA) and then analyzed for apoptosis using the terminal deoxynucleotidyltransferase-mediated uridine triphosphate end-labeling assay. Data points represent mean values ± standard deviations from three independent experiments. *: P < 0.05 (as determined by the Student t test) for the difference between the experimental group and the control group.

Table 3. HDACIs: Proportion of Cells Undergoing Apoptosis in Ovarian Carcinoma Cell Linesa
Cell lineControlSAHA (5 × 10−6 M)VPA (5 × 10−3 M)TSA (3 × 10−7 M)NaB (5 × 10−3 M)
  • HDACI: histone deacetylase inhibitors; SAHA: suberoylanilide bishydroxamine; VPA: valproic acid; TSA: trichostatin A; NaB: sodium butyrate.

  • a

    Apoptotic activity was measured in triplicate using the terminal deoxynucleotidyl transferase-mediated uridine triphosphate end-labeling (TUNEL) assay. Numbers indicate mean percentages ± standard deviations.

SK-OV-35 ± 256 ± 2249 ± 1762 ± 1843 ± 18
OVCAR-33 ± 139 ± 1536 ± 2049 ± 2241 ± 22
TOV-21G040 ± 1939 ± 2244 ± 2420 ± 8
OV-9010 ± 334 ± 926 ± 1035 ± 919 ± 6
TOV-112D4 ± 245 ± 1335 ± 1139 ± 1229 ± 11
OVCA4202 ± 042 ± 1531 ± 945 ± 1834 ± 9
OVCA429039 ± 740 ± 2148 ± 1230 ± 7
OVCA4321 ± 046 ± 638 ± 952 ± 2536 ± 12
OVCA4333 ± 249 ± 1828 ± 1044 ± 833 ± 16

Effect of Histone Deacetylase Inhibitors on Acetylation of Histones

Treatment of SK-OV-3 ovarian carcinoma cells with SAHA dramatically increased levels of acetylated H3 and H4 (Fig. 4A). This finding was representative of all HDACIs tested (data not shown).

Figure 4.

Protein levels, including levels of acetylated H3 and H4, in SK-OV-3 cells as measured by Western blot analysis. SK-OV-3 cells were treated with suberoylanilide bishydroxamine (SAHA), and cell lysates were harvested after 72 hours. Western blot analysis was performed using a series of antibodies. The amount of protein present was normalized relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) levels, and control cells were treated with vehicle alone. (A) Acetylated histone H3 and acetylated histone H4. (B) p21WAF1, p27KIP1, Bcl-2, Bax, cyclin D1, cyclin D2, and E-cadherin. (C) Caspase-9, caspase-3, and poly adenosine diphosphate-ribose polymerase (PARP). SK-OV-3 cells were treated with SAHA (5 × 10−6 M) for 24 hours. Before treatment (0 hours), only inactive procaspases and noncleaved PARP (112 kilodaltons [kD]) were detected. After treatment (24 hours), cleaved (i.e., activated) caspase fragments (p17 for caspase-9 and p10 for caspase-3) and cleaved PARP (85 kD) were detected.

Effect of Histone Deacetylase Inhibitors on the Expression of Cell Cycle– and Apoptosis-Related Proteins and E-Cadherin

p21WAF1 and p27KIP1 are cyclin-dependent kinase inhibitors (CDKIs) that bind to cyclin-dependent kinase complexes and decrease kinase activity, and may act as key regulators of the G0/G1 accumulation.15 We examined the effect of HDACIs on the expression of p21WAF1 and p27KIP1 in SK-OV-3 cells by Western blot analysis (SAHA: Fig. 4B; VPA, TSA, and NaB: data not shown). HDACIs markedly up-regulated levels of p21WAF1 and p27KIP1, which were expressed at negligible levels in untreated SK-OV-3 ovarian carcinoma cells. In contrast, HDACIs led to a reduction in cyclin D1 and cyclin D2 levels. In the same cell line, treatment with SAHA (5 × 10−6 M) reduced Bcl-2 levels by 70%, treatment with VPA reduced the expression of this protein by 25%, treatment with TSA reduced its expression by 25%, and treatment with NaB led to a 30% decrease in expression, whereas Bax expression remained unchanged (Fig. 4B).

We also analyzed the expression of selected caspase proteins, including caspase-9 and caspase-3, along with the expression of PARP, a substrate of caspase-3. After treatment with SAHA (5 × 10−6 M) for 24 hours, cleaved fragments of procaspase-9 and procaspase-3 (p17 and p10, respectively), as well as a cleaved fragment of PARP (85 kilodaltons), were detected (Fig. 4C), suggesting that the apoptosome pathway was activated by this HDACI.16

E-cadherin binds to β-catenin and can function as a tumor suppressor. The E-cadherin gene promoter contains CpG islands, which are frequently methylated in certain malignancies. HDACIs markedly increased the expression of E-cadherin in SK-OV-3 cells (SAHA: Fig. 4B; VPA, TSA, and NaB: data not shown).

Chromatin Immunoprecipitation

The effect of HDAC inhibition on the acetylation of histones H3 and H4, which are associated with the p21WAF1 gene promoter, was examined using ChIP. Chromatin fragments from SK-OV-3 cells cultured either with or without SAHA (5 ×10−6 M) for 3 days were immunoprecipitated with an antibody against acetylated H3 and H4. DNA from the immunoprecipitate was isolated, and a 255–base pair fragment of the p21WAF1 promoter region was amplified (Fig. 5). Approximately 12 times more p21WAF1 promoter DNA was associated with highly acetylated H3 and H4 histones in SAHA-treated cells compared with untreated cells (Fig. 5).

Figure 5.

Suberoylanilide bishydroxamine (SAHA) induces the accumulation of acetylated histones H3 and H4 (Ac-H3 and Ac-H4) in chromatin associated with the p21WAF1 gene. Soluble chromatin was immunoprecipitated with anti-acetylated histone H3/H4 antibodies from SK-OV-3 cells cultured in the presence or absence of 5 × 10−6 M SAHA for 3 days. Polymerase chain reaction primers for the p21WAF1 and β-actin gene promoter regions were used to amplify the DNA isolated from the immunoprecipitated chromatin.

Antitumor Activity In Vivo

We tested the ability of VPA to inhibit the growth of human SK-OV-3 ovarian tumors in immunodeficient mice over the course of 5 weeks of therapy. SK-OV-3 cells (5 × 106) led to the development of robust tumors in vivo (control mice). As shown in Figure 6A, administration of VPA remarkably suppressed the growth of these tumors. All treatment groups had statistically significantly smaller tumors than did the diluent control groups (P < 0.01). Tumor volume was measured at various time points throughout treatment, and at the termination of the study, tumors were carefully dissected and weighed. Our findings with regard to weight paralleled our volume measurements (Fig. 6B). Tumor weights were 80% lower in the treatment groups than in the control cohort (P < 0.01). During the study, all mice were weighed once weekly. Average body weights in the treatment groups were 90–103% of the average body weight in the control group (data not shown). In general, across all cohorts, all mice appeared to be healthy. No significant differences in mean weight, histology of internal organs, or mean blood chemistry values (including liver and hematopoietic parameters) were found between diluent-treated mice and those that received 5 weeks of therapy (data not shown). Histologic analysis of SK-OV-3 tumors from untreated mice revealed moderately differentiated carcinomas with small foci of necrosis and fibrosis, which constituted approximately 20% of the areas of the tumor sections. Approximately 50–60% of each of the tumor sections from mice treated with VPA exhibited necrosis and histologic changes suggestive of apoptosis, including the formation of apoptotic bodies. Fibrosis accounted for approximately 30% of the tumor area (data not shown). SK-OV-3 tumors were sampled for expression of p21WAF1 using immunohistochemical analysis of formalin-fixed, paraffin-embedded sections. SK-OV-3 ovarian carcinoma cells treated with VPA exhibited strong nuclear staining (Fig. 6C). Control SK-OV-3 ovarian carcinoma cells from untreated mice had negative or focal weak staining for p21WAF1 (Fig. 6D).

Figure 6.

Volume and weight of SK-OV-3 tumors in beige/nude/xid (BNX) mice receiving valproic acid (VPA). SK-OV-3 cells (5 × 106) were bilaterally and subcutaneously injected into the flanks of the BNX nude mice, leading to the formation of 2 tumors per mouse. Mice were divided randomly into control and experimental groups. VPA (10 mg per mouse) or diluent (control) was administered intraperitoneally 5 days per week for 5 weeks. (A) Time course of tumor volumes. Tumor volumes were measured each week and calculated as the product of the length, width, and height (see Materials and Methods) of each tumor. The mean volume ± standard deviation for the 10 tumors in each group is shown. Tumor volumes were significantly different between the experimental and control groups (P < 0.01). (B) Tumor weight at autopsy. After 5 weeks of therapy, tumors were removed from each group and weighed. Tumor weights were significantly different between the two groups (P < 0.01). (C,D) Immunohistochemical analysis of p21WAF1 expression in tumor tissue sections. (C) Ovarian carcinoma cells from mice treated with VPA show strong nuclear staining for p21WAF1. (D) Control ovarian carcinoma cells with either negative or focally weak staining for p21WAF1. Hematoxylin counterstain (C, D). Original magnification ×400 (C, D).

DISCUSSION

HDACIs can prevent proliferation and induce differentiation in numerous transformed cell types, including neuroblastoma, erythroleukemia, acute myelogenous leukemia, and carcinomas of the skin, breast, prostate, bladder, lung, colon, and cervix.17 The effect of HDACIs on ovarian carcinoma, however, has not been examined fully. A recent study by Terao et al.18 indicated that NaB had a significant growth-suppressing effect on human endometrial and ovarian carcinoma cells, irrespective of their p53 gene status.18 This finding led us to examine the effects of a wide array of HDACIs (SAHA, VPA, TSA, and NaB) on nine ovarian carcinoma cell lines (SK-OV-3, OVCAR-3, TOV-21G, OV-90, TOV-112D, OVCA420, OVCA429, OVCA432, and OVCA433) that express a mutated, nonfunctional form of the p53 protein.

We have demonstrated that each of the HDACIs is highly effective in suppressing the growth of human ovarian carcinoma cells. The prominent arrest of malignant cells in the G0/G1 phase of the cell cycle is likely to account for this effect. Expression of p21WAF1 and p27KIP1, two CDKIs that play important roles in blocking the cell cycle in the G1 phase,15 increased after treatment of ovarian carcinoma cells with HDACIs, supporting the idea that these proteins mediate a mechanism by which HDACIs inhibit ovarian carcinoma growth.

The enhanced expression of p21WAF1 was accompanied by an accumulation of acetylated histones H3 and H4, which are associated with the p21WAF1 gene. We and others have shown that HDACIs decrease levels of cyclin D1 and D2, and this decrease appears to occur at the transcriptional level.19, 20 Thus, HDACIs lead to decreased expression of D cyclins and increased expression of p21WAF1, with these effects probably combining to modulate the activity of the downstream pRb/E2F axis, thereby triggering cell cycle arrest.21

We have shown that treatment with HDACIs dramatically and significantly increases the number of apoptotic cells in all nine ovarian carcinoma cell lines studied. This effect was associated with a decrease in levels of the antiapoptotic protein Bcl-2. Experiments have shown that increased apoptotic activity is associated with the mitochondrial release of cytochrome C, which binds to and activates Apaf1. Activated Apaf1 and caspase-9 generate a complex, known as the apoptosome,16 that activates caspase-3, which cleaves many substrates associated with apoptosis, including PARP.16 Our findings of caspase-9 and caspase-3 activation and PARP cleavage suggest that this apoptosome pathway was activated after treatment with HDACIs.

The E-cadherin gene is involved in cell–cell adhesion, and the loss of E-cadherin protein function has been associated with enhanced metastatic growth of tumor cells.22 Inactivation of this gene by hypermethylation has been observed in breast carcinoma cells and in primary breast tumors.23 We found that transcription of E-cadherin was up-regulated in ovarian carcinoma cells treated with HDACIs, suggesting a gain of tumor suppressor function in response to the inhibition of HDAC.

Although HDACIs have been shown to have an antiproliferative effect in vitro, a number of limitations hamper their clinical use. NaB, a low-potency HDACI, has been studied extensively. It possesses antitumor activity and can induce differentiation in some malignant cell lines, but its clinical utility has been restricted by its short half-life (5 minutes), which limits its ability to reach therapeutic levels in plasma. TSA is of limited therapeutic use because of toxic side effects in vivo.

SAHA and VPA, however, are relatively safe and nontoxic in vivo. For this reason, we focused on these two HDACIs in the current study. SAHA is currently in Phase I of clinical trials for the treatment of solid and hematologic tumors, and VPA has been used in the treatment of epilepsy for almost 30 years. Our in vitro studies show that VPA at a concentration of 0.3–1.5 mM inhibited cell proliferation, induced cell cycle arrest, and stimulated apoptosis in SK-OV-3 ovarian carcinoma cells. In patients with epilepsy, this range of VPA concentrations can be achieved in serum through the administration of a daily dose of 20–30 mg/kg. These data also are consistent with our in vivo data. Because we did not analyze the pharmacokinetics of this HDACI in our murine model, we could not conclude that in vivo levels of this HDACI were comparable to those used in our in vitro studies. Furthermore, the binding of HDACIs to plasma proteins in vivo could differ from the binding of HDACIs to the proteins present in bovine calf serum (which was used for in vitro experiments), and this difference could be accompanied by a difference in efficacy.

Our histologic data showed that the SK-OV-3 cells formed moderately differentiated adenocarcinomas in nude mice. Extensive necrosis, apoptosis, and fibrosis involving approximately 30–60% of the tumor area was observed in mice treated with VPA, and this antitumor activity was not accompanied by any major side effects, raising the possibility that VPA may serve as a useful adjuvant therapy option for the treatment of ovarian carcinoma. Because we started HDACI treatment on the day after malignant cells were injected into mice, we cannot conclude that HDACIs are effective against bulky tumors, which are frequently observed in the clinical setting; thus, further studies are necessary to clarify this issue. Nonetheless, HDACIs have profound antitumor activity in vivo, and we have confirmed that a small tumor burden can be eradicated by these compounds in mice. This finding suggests that HDACI therapy may be particularly effective for individuals who have minimal residual disease after curative surgery, chemotherapy, and/or radiotherapy. Furthermore, VPA has convenient pharmacokinetic properties, with a significantly longer biologic half-life than the other HDACIs.24

In summary, HDACIs exhibit antiproliferative activity and potently induce apoptosis in human ovarian carcinoma cells. These events are accompanied by the induction of p21WAF1 and p27KIP1 and the down-regulation of several antiapoptotic and cell cycle–related proteins, such as Bcl-2, cyclin D1, and cyclin D2. Furthermore, VPA significantly inhibited tumor growth in nude mice without any apparent toxicity. These findings suggest that HDACIs may be particularly effective in the treatment of ovarian carcinoma.

Ancillary