The first 2 authors contributed equally to this article.
Decitabine and suberoylanilide hydroxamic acid (SAHA) inhibit growth of ovarian cancer cell lines and xenografts while inducing expression of imprinted tumor suppressor genes, apoptosis, G2/M arrest, and autophagy†
Article first published online: 16 MAR 2011
Copyright © 2011 American Cancer Society
Volume 117, Issue 19, pages 4424–4438, 1 October 2011
How to Cite
Chen, M.-Y., Liao, W. S.-L., Lu, Z., Bornmann, W. G., Hennessey, V., Washington, M. N., Rosner, G. L., Yu, Y., Ahmed, A. A. and Bast, R. C. (2011), Decitabine and suberoylanilide hydroxamic acid (SAHA) inhibit growth of ovarian cancer cell lines and xenografts while inducing expression of imprinted tumor suppressor genes, apoptosis, G2/M arrest, and autophagy. Cancer, 117: 4424–4438. doi: 10.1002/cncr.26073
The first 2 authors contributed equally to this article. We acknowledge support from the National Cancer Institute for our Translational Chemistry Core Facility.
- Issue published online: 16 SEP 2011
- Article first published online: 16 MAR 2011
- Manuscript Accepted: 2 FEB 2011
- Manuscript Revised: 21 JAN 2011
- Manuscript Received: 17 SEP 2010
- Anderson Cancer Center Specialized Programs of Research Excellence (SPORE)
- National Cancer Institute (NCI). Grant Number: P50 CA83639
- National Foundation for Cancer Research
- Zarrow Foundation
- NCI Cancer Center. Grant Number: CA016672
- ovarian cancer;
- suberoylanilide hydroxamic acid;
- imprinted tumor suppressor genes;
Epigenetic therapy has had a significant impact on the management of hematologic malignancies, but its role in the treatment of ovarian cancer remains to be defined. The authors previously demonstrated that treatment of ovarian and breast cancer cells with DNA methyltransferase and histone deacetylase (HDAC) inhibitors can up-regulate the expression of imprinted tumor suppressors. In this study, demethylating agents and HDAC inhibitors were tested for their ability to induce re-expression of tumor suppressor genes, inhibiting growth of ovarian cancer cells in culture and in xenografts.
Ovarian cancer cells (Hey and SKOv3) were treated with demethylating agents (5-aza-2′-deoxycytidine [DAC] or 5-azacitidine [AZA]) or with HDAC inhibitors (suberoylanilide hydroxamic acid [SAHA] or trichostatin A [TSA]) to determine their impact on cellular proliferation, cell cycle regulation, apoptosis, autophagy, and re-expression of 2 growth inhibitory imprinted tumor suppressor genes: guanosine triphosphate-binding Di-RAS-like 3 (ARHI) and paternally expressed 3 (PEG3). The in vivo activities of DAC and SAHA were assessed in a Hey xenograft model.
The combination of DAC and SAHA produced synergistic inhibition of Hey and SKOv3 cell growth by apoptosis and cell cycle arrest. DAC induced autophagy in Hey cells that was enhanced by SAHA. Treatment with both agents induced re-expression of ARHI and PEG3 in cultured cells and in xenografts, correlating with growth inhibition. Knockdown of ARHI decreased DAC-induced autophagy. DAC and SAHA inhibited the growth of Hey xenografts and induced autophagy in vivo.
A combination of DAC and SAHA inhibited ovarian cancer growth while inducing apoptosis, G2/M arrest, autophagy, and re-expression of imprinted tumor suppressor genes. Cancer 2011;. © 2011 American Cancer Society.
Ovarian cancer is the second most common gynecologic malignancy in the United States and has the highest fatality-to-case ratio.1 Despite advances in surgery and chemotherapy, the cure rate is <30% for advanced stage ovarian cancer,2 underscoring the urgent need for more effective therapy based on a better understanding of the biology of the disease.
Molecular mechanisms underlying the development and progression of ovarian cancer remain poorly understood. Genetic alterations have been detected in different ovarian cancers with activating mutations and amplification of oncogenes as well as inactivating mutations and loss of heterozygosity of tumor suppressor genes.3
Epigenetic changes also have been implicated in malignant transformation and progression of different cancers, including ovarian cancer.4 Aberrant DNA methylation in the CpG islands within promoter regions is associated with silencing of tumor suppressor genes, which is considered equivalent to the effect of inactivating mutations or deletion of tumor suppressor genes.5 Aberrant DNA methylation can be reversed with hypomethylating agents, such as 5-azacitidine (AZA), and 5-aza-2′-deoxycytidine (decitabine; DAC).5, 6 Re-expression of genes that inhibit growth, promote apoptosis, and induce autophagy could exert antitumor activity. Although significant therapeutic activity has been reported in patients with hematopoietic malignancies,7 few trials have been performed in solid tumors, including ovarian cancer. Recent studies with AZA suggest that re-expression of silenced genes can be achieved and that sensitivity to carboplatin can be restored in a fraction of patients with platinum-resistant ovarian cancer.8
Epigenetic regulation of tumor suppressor genes also occurs through histone modification. Acetylation of histones H3 and H4 is associated with active gene transcription, whereas deacetylation represses gene expression. These changes are catalyzed by histone acetyltransferases and histone deacetylases (HDAC). Several compounds, including suberoylanilide hydroxamic acid (SAHA) and trichostatin A (TSA), inhibit HDAC activity and can induce global and gene-specific hyperacetylation, enhancing the expression of numerous genes. It has been demonstrated that these compounds can inhibit cancer growth in different preclinical models, and they are being evaluated in clinical trials with SAHA (vorinostat) for treating patients with cutaneous T-cell lymphoma.9, 10 When it was used as a single agent in platinum-refractory ovarian cancer, SAHA was well tolerated but had minimal activity.11 A recent phase 1b-2a study to assess the effects of hypomethylating agent azacitidine in patients with platinum-resistant or platinum-refractory epithelial ovarian cancer indicated that treatment with a hypomethylating agent could partially reverse platinum resistance.8
Combined use of demethylating agents and HDAC inhibitors have achieved greater antitumor activity12, 13; however, the mechanisms underlying their activity are not well defined but may relate to the up-regulation of growth-inhibitory and proapoptotic genes that have been silenced during oncogenesis by hypermethylation and histone deacetylation. Although both growth-inhibitory and growth-stimulatory genes may be up-regulated by these inhibitors, it is believed that tumor progression during oncogenesis involves the silencing of tumor suppressor genes in excess of oncogenes. Thus, growth-inhibitory genes may be up-regulated preferentially by treatment with demethylating agents and HDAC inhibitors. Early clinical trials have demonstrated that demethylating agents and HDAC inhibitors can be administered together at effective doses to treat hematologic malignancies.13
Several putative tumor suppressor genes are imprinted and expressed from only a single, nonimprinted allele in normal cells. In our previous studies, we identified 7 imprinted tumor suppressor genes that are down-regulated in ovarian cancers.14 Among these genes, guanosine triphosphate-binding Di-RAS-like 3 (ARHI) and paternally expressed 3 (PEG3) are the most markedly down-regulated with reduced expression in approximately 63% to 88% and 75% of ovarian cancers, respectively.15, 16 ARHI and PEG3 can be down-regulated by loss of heterozygosity or can be silenced by promoter hypermethylation of both alleles.16 In addition, ARHI expression is down-regulated transcriptionally and post-transcriptionally in ovarian cancer cells.17 Re-expression of ARHI can be achieved in cancer cells by treatment with AZA and TSA that, in combination, can stimulate transcription from both the imprinted and nonimprinted alleles.18 Re-expression of ARHI and PEG3 strongly inhibit clonogenic growth of ovarian cancers.16 Re-expression of ARHI causes autophagic death of cultured human ovarian cancer cells and induces tumor dormancy in xenografts.19 Autophagy is characterized by the accumulation of multilamellar vacuoles that engulf cytoplasm and organelles, forming autophagosomes marked by microtubule-associated protein light chain 3 (MAP-LC3). Autophagosomes then fuse with lysosomes, releasing their contents for hydrolysis, which can temporarily sustain energy production by stressed cells but can prove lethal if prolonged.20
In the current study, we measured the antiproliferative activity of 2 demethylating agents (DAC and AZA) and 2 HDAC inhibitors (SAHA and TSA) individually and in combination against human Hey and SKOv3 ovarian cancer cells. To identify mechanisms underlying growth inhibition, we measured changes in cell cycle, apoptosis, and autophagy. We have tested the hypothesis that growth inhibition would be associated with re-expression of growth-inhibitory tumor suppressor genes, including ARHI and PEG3, and that these genes might be used as biomarkers for effective dosing of demethylating agents and HDAC inhibitors in cell culture and in xenografts.
MATERIALS AND METHODS
Reagents and Cells
AZA (Pharmion Corp., Boulder, Colo), and DAC (Sigma-Aldrich, St. Louis, Mo) were dissolved in distilled water. TSA (Sigma-Aldrich) was dissolved in ethanol. SAHA was synthesized according to Gediva et al21 and dissolved in dimethyl sulfoxide. The characterization of SAHA is in agreement with the reference, and the purity is >95% based on liquid chromatography mass spectrometry analysis. Cells were maintained in RPMI-1640 medium (Hey cells) or McCoy 5A medium (SKOv3 cells) supplemented with 10% fetal bovine serum, 100 mM L-glutamine, 100 μg/mL streptomycin, and 100 U/mL penicillin. The identity of both cell lines was confirmed by DNA satellite signature analysis.
Cell Growth Assays
The growth of Hey and SKOv3 cells was evaluated by sulforhodamin B (SRB) staining as described previously.19 Briefly, Hey cells (1.5 × 103) and SKOv3 cells (2 × 103) were plated in triplicate in 96-well plates for 24 hours and treated with DAC, AZA, SAHA, and TSA individually or in combination with serial, 2-fold dilutions of the 2 agents. Culture media were changed daily with freshly added inhibitors. Cells were harvested on Day 5 for SRB assays, and the values were normalized to untreated controls.
RNA Extraction and Real-Time Reverse Transcriptase-Polymerase Chain Reaction
Expression of ARHI and PEG3 was measured using real-time quantitative reverse transcriptase-polymerase chain reaction, as described previously.16 The primers for ARHI and PEG3 were purchased from Applied Biosystems (Foster City, Calif). Relative levels of ARHI and PEG3 messenger RNA (mRNA) were normalized to a concurrent determination for glyceraldehyde 3-phosphate dehydrogenase mRNA. All samples were measured in triplicate, and each experiment was repeated twice.
Measurement of Apoptosis Induction
Cultured cells were treated with DAC, SAHA, or both agents combined for 5 consecutive days without changing the medium. Caspase 3 and 7 activities were measured using the Caspase-Glo 3 of 7 Assay (Promega, Madison, Wis).
Cell Cycle Analysis
Cells were grown for 24 hours and treated with DAC, SAHA, or their combination for 1 day, 3 days, or 5 days. Cell viability was assessed by staining with propidium iodide (PI) and was analyzed by flow cytometry.
Detection of Autophagy With Green Fluorescence Protein-Tagged MAP-LC3
Cultured cells were treated with DAC, SAHA, or both agents combined for 3 days. On Day 3, the cells were transfected with green fluorescence protein (GFP)-tagged MAP-LC3 (GFP-LC3) plasmid. After 24 hours, the cells were fixed in 4% paraformaldehyde for 30 minutes and mounted for confocal microscopy. GFP fluorescence was observed under a confocal microscope, and autophagic cells that demonstrated punctate GFP-LC3 staining were counted.
Detection of Autophagy With Acridine Orange Staining
Ovarian cancer cells were seeded and treated as described above for the cell cycle assays. At the indicated times, the cells were incubated with medium containing 0.5 μg/mL acridine orange (Molecular Probes, Eugene, Ore) for 15 minutes in the dark, detached by trypsinization, washed, and analyzed by fluorescence-activating cell sorting analysis.
Transmission Electron Microscopy
Transmission electron microscopy (TEM) was used to confirm morphologically the induction of autophagy by examining alterations in the subcellular structures of Hey cells in culture and in xenografts. Cells were cultured and treated with DAC and/or SAHA for 5 days. Cultured cells and tumor xenografts were fixed and prepared for TEM as described previously.19 Representative areas were chosen for ultrathin sectioning and viewed with a JEM 1010 transmission electron microscope (JEOL Ltd., Tokyo, Japan).
Small Interfering RNA Transfection
Hey cells were cultured and treated with DAC and/or SAHA for 3 days. Transfection with control or ARHI small interfering RNA (siRNA) was performed on Day 2 using the DharmaFECT-4 reagent (Dharmacon Research, Lafayette, Colo). A mixture of siRNA (100 nM final concentration) and transfection reagents was incubated at room temperature for 20 minutes and then added to the cells. After incubation for 48 hours, the cells were harvested for confocal microscopy with GFP-LC3 or flow cytometry with acridine orange staining.
Growth of Ovarian Cancer Xenografts in Nude Mice
In vivo studies with Hey ovarian cancer xenografts were carried out in 6-week-old female Balb-c nu/nu mice according to protocols approved by the Institutional Animal Care and Use Committee at the University of Texas MD Anderson Cancer Center. Twenty mice were divided into 4 treatment groups. Twelve additional mice were divided into 4 groups for histologic studies on Day 22. All mice were injected intraperitoneally with 1 × 106 Hey cells in 200 μL RPMI-1640 media. Two days later, the mice were injected intraperitoneally with 1) vehicle (control group), 2) DAC (0.8 mg/kg/3 times weekly), 3) SAHA (12.5 mg/kg/5 times weekly), or 4) a combination of DAC and SAHA at the same dose as the single-agent treatments. Treatment was continued for 21 days. On Day 22, tumors from 3 mice of each group were collected and prepared for TEM. The remaining 5 mice from each group were evaluated daily for morbidity and mortality.
All experiments were repeated independently at least twice. To assess in vitro growth inhibition and synergistic inhibition, we used the SYNERGY22 program, which estimates dose-response curves for each agent, either alone or combined, and quantifies drug interaction at different inhibitory levels. One may conclude synergy when the interaction index is <1 and its 95% confidence interval is <1. The analysis is based on the median-effect principle and the combination index method.23 Regression-based analysis was used to test for associations between growth inhibition and gene expression. For in vivo studies, the statistical significance of differences in survival between mice was analyzed using the log-rank test.24 We used the Efron method25 to handle observations that had tied survival times. P values were obtained by 1-sided analysis, and significance was assumed at P < .05.
DNA Methyltransferase Inhibitors and HDAC Inhibitors Decrease Growth of Ovarian Cancer Cells
To identify optimal drugs, we compared the antiproliferative activity of 2 DNA methylation inhibitors (DAC and AZA) and 2 HDAC inhibitors (SAHA and TSA) in 2 human ovarian cancer cell lines. Hey cells and SKOv3 cells were treated with increasing concentrations of DAC (0-100 μM), AZA (0-100 μM), SAHA (0-32 μM), or TSA (0-2 μM). Growth was measured after 5 days of treatment. The viability of both Hey cells and SKOv3 cells was reduced in a dose-dependent manner for all 4 agents (Fig. 1). For cells that were treated with DAC, the reduction in viability at increasing doses could be fitted by hyperbolic curves, which decreased to asymptotes at approximately 30% for Hey cells and 60% for SKOv3 cells. In SKOv3 cells, there was no further increase in the antiproliferative activity of DAC at concentrations >32 μM. However, treatment with AZA at doses >50 μM killed all Hey cells and decreased SKOv3 cells at an asymptote of 15%. In contrast to the DNA methylation inhibitors, both HDAC inhibitors were very toxic to cells, and all cells were killed at concentrations >8 μM and >0.72 μM for SAHA and TSA, respectively (Fig. 1). The 50% inhibitory concentration (IC50) and the IC75 for each agent, individually and in combination, are presented in Table 1.
|Hey Cells||SKOv3 Cells|
|Treatment||IC50, μM||IC75, μM||IC50, μM||IC75, μM|
DAC and SAHA Exert Synergistic Inhibition of Cell Growth
To explore the potential interaction between DNA methylation inhibitors and HDAC inhibitors, we treated Hey cells and SKOv3 cells with DAC or AZA in combination with SAHA or TSA. The concentrations for each agent were determined based on the combination design proposed by Chou,26 in which a mixture of 1 DNA methylation inhibitor and 1 HDAC inhibitor with the diagonal constant ratio ([IC50 of DAC or AZA]:[IC50 of SAHA or TSA]) was serially diluted before addition to the cells. In every set of experiments, the combination of 1 DNA methylation inhibitor and 1 HDAC inhibitor resulted in a significantly more pronounced cell kill compared with treatment using each single agent (Fig. 1, Table 1). When results for each combination of inhibitors were analyzed for additive, synergistic, or antagonistic effects, the most prominent synergistic interaction was observed between DAC and SAHA for both Hey and SKOv3 cells. Other combinations resulted in primarily additive toxicity. On the basis of these results, we chose the combination of DAC and SAHA at their IC50 values for all subsequent cell culture studies.
Treatment With DAC and SAHA Induces Apoptosis
To examine whether the antiproliferative effect of DAC and SAHA could be attributed to the induction of apoptosis, caspase 3/7 activities were analyzed in Hey and SKOv3 cells that had been treated for 5 days with DAC and/or SAHA at concentrations that resulted in 50% inhibition of cell growth. Treatment with DAC and SAHA alone induced caspase activity in SKOv3 cells but not in Hey cells, as illustrated in Figure 2. However, the combination of DAC and SAHA increased caspase activity synergistically in both cell lines.
A Combination of DAC and SAHA Arrests Ovarian Cancer Cells in G2/M
To determine whether the antiproliferative effect of DAC and SAHA also might result from cell cycle arrest, cell cycle analyses were performed on cells that were treated for 1 day, 3 days, and 5 days. The percentages of cells in G1, S, and G2/M phase are illustrated in Figure 3. When used individually, treatment with SAHA had only a modest and inconsistent effect on the cell cycle over the course of the experiment. DAC arrested a progressively greater number of cells in G2/M over time with 16.1% of cells arrested on Day 1, 28.7% arrested on Day 3, and 35.3% arrested on Day 5 in Hey cells. DAC alone failed to affect the fraction of SKOv3 cells in G2/M, but the combination of the 2 agents produced G2/M arrest in both cell lines. In Hey cells, treatment with a combination of DAC and SAHA increased the fraction of cells in G2/M progressively from 23.8% on Day 1 to 44.9% on Day 5. A more modest effect was observed with SKOv3 cells on Day 5, in which 20.4% of cells that were treated with the combination were in G2/M compared with 14.1% for the controls. Although both cell lines responded to the combination treatment by increasing the fraction of cells in G2/M phase, the change in Hey cells was more evident. The underlying mechanism that contributes to this dichotomous effect is not clear. We also quantified the sub-G0/G1 fraction as an additional assessment of apoptosis. Consistent with the caspase activity, DAC alone produced a modest increase in the fraction of sub-G0/G1 cells, whereas SAHA had no effect. Combination treatment with DAC and SAHA resulted in a synergistic increase in apoptotic cells (Fig. 2C,D).
DAC and SAHA Induce Autophagy in Hey Ovarian Cancer Cells
DAC and SAHA inhibited the proliferation of both Hey and SKOv3 cells by apoptosis and G2/M arrest. A third possible mechanism that could contribute to their growth inhibitory effects is autophagic cell death. To assess this possibility, Hey and SKOv3 cells were transiently transfected with a GFP-LC3 plasmid, treated with DAC and SAHA for 3 days, and then examined under a confocal fluorescence microscope. Treatment of Hey cells with DAC induced autophagy, as indicated by an increase in GFP-LC3 puncta (Fig. 4A). Although SAHA alone did not induce autophagy, it enhanced the ability of DAC to induce autophagy. When the GFP-LC3 puncta were counted in Hey cells, the number of puncta increased from 60 puncta per 100 cells for the control cells to 150 puncta per 100 cells for the cells treated with combination of DAC and SAHA. In contrast, autophagy was not observed in SKOv3 cells after treatment with these agents.
To confirm that these agents induced autophagy, we used acridine orange staining followed by flow cytometry. Figure 4B indicates that DAC induced autophagy in Hey cells with 36% acridine-positive cells on Day 3 and 35% acridine-positive cells on Day 5. Consistent with the GFP-LC3–transfected cells, SAHA alone did not induce autophagy, but the combination of SAHA and DAC increased acridine-positive cells to 49% on Day 3 and to 51% on Day 5. No autophagy was observed for SKOv3 cells in any treatment group. Finally, we used TEM to document autophagy in Hey cells that were treated for 5 days with diluent, DAC, SAHA, or both agents. Figure 4C indicates that treatment with DAC induced the formation of autophagosomes compared with the control group. Whereas SAHA alone did not induce autophagy, the addition of DAC + SAHA further enhanced the number of autophagosomes.
Up-Regulation of ARHI and PEG3 Correlates With Growth Inhibition Induced by a Combination of DAC and SAHA
We demonstrated previously that ARHI and PEG3 are down-regulated in the majority of ovarian cancers and that their expression can be increased by treatment with demethylating agents and HDAC inhibitors.16 In the current study, we evaluated whether the induction of ARHI and PEG3 in ovarian cancer cells was correlated with the inhibition of growth by DAC and SAHA. The expression of ARHI and PEG3 in Hey and SKOv3 cells was quantified after treatment for 5 days with DAC, SAHA, or both agents combined (Fig. 5). Cell growth was measured in parallel cultures.
Treatment of Hey cells with DAC dramatically induced ARHI and PEG3 expression and was associated with a decrease in cell growth (Fig. 5A). Treatment with SAHA alone had little effect on ARHI or PEG3 expression but completely inhibited cell growth. The combination of DAC and SAHA was more potent for inhibiting cell growth and produced a dose-dependent increase in ARHI and in PEG3 mRNA. These data indicate that levels of ARHI and PEG3 are correlated inversely with growth inhibition produced by a combination of DAC and SAHA.
For SKOv3 cells, DAC induced PEG3 and, to a lesser extent, ARHI but had little effect on cancer cell growth (Fig. 5B). Like what we observed in Hey cells, SAHA alone failed to induce ARHI or PEG3 expression in SKOv3 cells but inhibited growth by >90%. A combination of DAC plus SAHA produced dose-dependent induction of ARHI and PEG3 and >90% inhibition of cancer cell growth.
A regression-based analysis was used test for an association between growth inhibition and gene expression. We fit a regression model to the gene expression data, adjusting for cell line, drug, and gene. Association was determined by significance in the regression slope coefficients. Significance was assumed at P < .05. Growth inhibition of Hey cells was correlated with the induction of ARHI after treatment with DAC (P = .000004) or with a combination of DAC and SAHA (P = .00002). PEG3 levels appeared to vary inversely with growth inhibition by DAC or a combination of DAC and SAHA, but a slight depression in PEG3 levels at IC75 precluded a statistically significant correlation in each case. Growth inhibition by DAC or a combination of DAC and SAHA in SKOv3 cells was correlated with ARHI (P = .03 and P = .01, respectively) and with PEG3 levels (P = .00009 and P = 2.21E-13, respectively).
Expression of ARHI Is Required for Optimal Induction of Autophagy by a Combination of DAC and SAHA
Re-expression of ARHI by transfection or with a Tet-inducible promoter can produce autophagy in SKOv3 and Hey cells.19 The dramatic induction of ARHI after DAC treatment and the known role of ARHI in autophagy raise the possibility that DAC-induced autophagy may be mediated through ARHI. To test this hypothesis, Hey cells were transfected with siARHI before treatment with DAC and SAHA. Induction of autophagy was quantified by acridine orange staining and flow cytometry. Knockdown of ARHI expression dramatically reduced autophagy induced by DAC or DAC + SAHA, as indicated in Figure 6.
To examine whether growth inhibition produced by DAC and SAHA also may depend, at least in part, on the re-expression of ARHI and PEG3, Hey and SKOv3 cells were transfected with control siRNA or with a mixture of ARHI siRNA and PEG3 siRNA before they were treated with a combination of DAC and SAHA. Changes in cell proliferation were measured by SRB analysis. Our results demonstrated that knockdown of ARHI and PEG3 had no effect on DAC-induced or SAHA-induced growth inhibition (data not shown), suggesting that the observed growth inhibition was mediated by other DAC-responsive and SAHA-responsive genes.
DAC and SAHA Exhibit Additive Inhibition of Ovarian Cancer Xenograft Growth
To further validate our in vitro studies, we used nude mice with intraperitoneal human Hey ovarian cancer xenografts. Groups of mice were treated intraperitoneally with 1 of the 4 treatments described above (see Materials and Methods). The effect of treatment on survival was evaluated, and the presence of autophagosomes was examined by TEM. Figure 7A displays the Kaplan-Meier survival curves for each of the 4 treatment groups. A test of significance of each difference from the control was provided by the log-rank test using the Efron method25 to resolve observations that had tied survival times. Significance was assumed at P < .05. Although SAHA alone did not improve survival (P = .135), treatment with DAC alone (P = .0169) or DAC combined with SAHA (P = .00184) for 3 weeks significantly increased survival.
Autophagy Is Observed in Ovarian Cancer Xenografts After DAC and SAHA Treatment
To determine whether the action of DAC and SAHA in vivo was associated with the induction of autophagy, 3 mice in each group were killed after 21 days of treatment, and their xenograft tumors were harvested for TEM to examine for the presence of autophagosomes. Although no or a few autophagosomes were detected in the control and SAHA-treated groups, the number of autophagosomes was significantly increased in the DAC-treated groups, and the group that was treated with a combination of DAC and SAHA had the greatest number of autophagosomes (Fig. 7B). Thus, the additive inhibition of xenograft growth observed with DAC and SAHA may be caused in part by the autophagic death of ovarian cancer cells.
Treatment With DAC and SAHA Reactivates the Expression of ARHI and PEG3 in Hey Ovarian Cancer Xenografts
The combined treatment of DAC + SAHA reactivated both ARHI and PEG3 expression in cultured cells (Fig. 5), as discussed above. To determine whether the combined treatment also could reactivate the expression of ARHI and PEG3 in xenografts, we quantified ARHI and PEG3 transcripts in xenografts from mice that were treated with DAC + SAHA and compared the results with those from control mice. Both ARHI and PEG3 transcripts were increased by approximately 1.5-fold in xenografts from DAC + SAHA-treated mice, as indicated in Figure 8. The increase in ARHI expression was demonstrated further by increased immunohistochemical staining with anti-ARHI antibody.
Our current study demonstrates that a combination of a demethylating agent and a HDAC inhibitor can exert synergistic inhibition of growth in ovarian cancer. This principle may apply to a variety of cancers that arise from different sites.27-30 Previous reports from the ovarian cancer literature have documented that demethylating agents31 and histone deacetylating agents32, 33 can inhibit the growth of multiple cell lines and primary cultures of ascites tumor cells,32 enhancing the activity of carboplatin31 or paclitaxel.32, 33 Moreover, treatment of xenografts with DAC and the HDAC inhibitor belinostat enhanced the activity of cisplatin, which is associated with the up-regulation of human mutL homolog 1 (hMLH1).34 No interaction was noted in that study between DAC and belinostat in the absence of cisplatin. In our study, maximally tolerated doses of DAC significantly inhibited Hey xenograft growth, and treatment with DAC in combination with SAHA provided superadditive inhibition of xenograft growth that approached statistical significance (P = .068). Differences in outcome between cell cultures and xenografts may relate to pharmacologic and pharmacokinetic considerations, particularly for SAHA. For patients with myelodysplastic syndrome35, 36 or chronic myelogenous leukemia,37 the administration of low-dose DAC as an intravenous infusion has produced therapeutic responses. However, we chose to use maximally tolerated doses of DAC because of the route of administration (intraperitoneal vs intravenous) and because we wanted to maximize the inhibitory effect of DAC and SAHA in the growth of Hey tumor xenografts. The mice appeared to tolerate these doses and regimens, because there were no early deaths or significant weight losses with the combination treatment.
The growth inhibition produced in cell cultures and xenografts with demethylating agents and HDAC inhibitors may relate to the re-expression of growth-inhibitory tumor suppressor genes. Although many genes may be up-regulated or down-regulated when cells are treated with demethylating agents alone38, 39 or in combination with HDAC inhibitors,40 a combination of the 2 agents can facilitate the re-expression of hypermethylated40 and/or imprinted18 tumor suppressor genes, including ARHI and PEG3. Treatment with DAC increased the expression of ARHI and PEG3 in both cell lines. Although treatment with SAHA alone had minimal effect on ARHI and PEG3 expression, the combination of DAC and SAHA had a greater effect on the induction of both ARHI and PEG3. Growth inhibition produced by the combined treatment correlated inversely with levels of ARHI and PEG3.
The mechanisms by which DAC and SAHA inhibit tumor growth are multiple and may vary between different ovarian cancers. Both ARHI and PEG3 can inhibit clonogenic growth, and ARHI produces cell cycle arrest.16 In the current study, the combination of DAC and SAHA produced apoptosis in both Hey and SKOv3 cell lines. In previous reports, treatment with SAHA has produced apoptosis in ovarian cancer cells,32, 33 but, to our knowledge, the induction of autophagy has not been noted previously. Treatment with DAC induced autophagy in Hey ovarian cancer cells, and autophagy could be enhanced by concomitant treatment with both DAC and SAHA. Because it has been demonstrated that ARHI can induce autophagy,19 increased levels of this gene may contribute to autophagic cell death. Knockdown of ARHI reduced autophagy, consistent with the importance of ARHI in regulating this function.
Taken together, our data suggest that a combination of a demethylating agent and an HDAC inhibitor should have greater impact on ovarian cancer than either treatment alone, and this finding deserves further evaluation in the clinic. Re-expression of ARHI and PEG3 may provide useful biomarkers for identifying biologically optimal doses of the inhibitors.
This work was supported in part with funds from The University of Texas MD Anderson Cancer Center Specialized Programs of Research Excellence (SPORE) in Ovarian Cancer from the National Cancer Institute (NCI) (P50 CA83639), the National Foundation for Cancer Research, the Zarrow Foundation, and Mr. Stuart Zarrow. The University of Texas MD Anderson Cancer Center Translational Chemistry Core Facility also is supported by NCI Cancer Center Support Grant CA016672.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
- 1Ovarian cancer. In: Hong WK, Bast RC Jr, Hait W, et al, eds. Holland-Frei Cancer Medicine. 8th ed. American Association for Cancer Research. Shelton CT: People's Medical Publishing House-USA; 2010: 1344-1375., , .
- 8A phase IB-IIA study to reverse platinum resistance by use of a hypomethylating agent azacitidine in platinum-resistant or refractory epithelial ovarian Cancer. In press., , , et al.
- 22http://biostatistics.mdanderson.org/SoftwareDownload. Accessed on September 1, 2010., , , . SYNERGY Software.
- 23The median-effect principle and the combination index for quantification of synergism and antagonism. In: Chou TC, Rideout DC, eds. Synergism and Antagonism in Chemotherapy. San Diego, CA: Academic Press; 1991: 61-101..