Sensitization of ovarian carcinoma cells to the atypical retinoid ST1926 by the histone deacetylase inhibitor, RC307: enhanced DNA damage response

Abstract

The synthetic atypical retinoids containing an adamantyl group exhibit antiproliferative or proapoptotic activities. Apoptosis induction is a dose-dependent effect independent of retinoid receptors. We have reported that induction of apoptosis by the atypical retinoid, ST1926, is associated with early manifestations of genotoxic stress. Indeed, in this study performed in ovarian carcinoma cells, we show that exposure to ST1926 resulted in an increase of early markers of DNA damage, including ATM and H2AX phosphorylation. In addition, we found that a novel histone deacetylase (HDAC) inhibitor (RC307) was able to enhance sensitivity of ovarian carcinoma cells to ST1926. Under conditions where single-agent treatment caused only antiproliferative effects, the combination of the atypical retinoid and HDAC inhibitor resulted in marked apoptotic cell death with a more rapid onset in wild-type p53 ovarian carcinoma cells. The sensitization to ST1926-induced apoptosis was associated with an enhanced DNA damage response, because a prolonged expression of DNA damage markers (e.g., H2AX, p53 and RPA-2 phosphorylation) and a marked activation of DNA damage checkpoint kinases (in particular, phosphorylation of Chk1) were observed indicating an accumulation of DNA damage by the ST1926/HDAC inhibitor combination. The study provides additional support to the role of DNA damage as a primary event leading to the activation of apoptosis in ovarian carcinoma cells by adamantyl retinoids and documents the potential therapeutic efficacy of the combination of ST1926 and HDAC inhibitors of the novel series.

Atypical retinoids, also known as retinoid-related molecules, are synthetic compounds characterized by a structure reminiscent of classical retinoids.1, 2 The best-known compounds of this series are the adamantyl retinoids, ST1926 and CD437, which are potent inducers of apoptosis with promising potential in antitumor therapy.2, 3 ST1926 is undergoing phase I clinical trial in ovarian carcinoma.

The proapoptotic and cytotoxic activities of adamantyl retinoids do not appear to be mediated by retinoid receptors.4 The molecular mechanisms involved in induction of apoptosis are not clearly defined and are still matter of debate. Several lines of evidence support the involvement of a genotoxic stress in the mechanism of apoptosis induction. In particular, the early activation or modulation of several factors implicated in DNA damage response suggests that DNA damage is a primary event in the mechanism of cytotoxic activity of adamantyl retinoids.5–8 Based on this putative mechanism of action, this study was designed to explore the ability of an histone deacetylase (HDAC) inhibitor (RC307, Fig. 1) to influence the sensitivity of ovarian carcinoma cells to ST1926. Indeed, because histone acetylation modulates chromatin structure and gene expression, HDAC inhibition could influence DNA damage response. A number of HDAC inhibitors, including hydroxamic acid-based compounds, are known to sensitize tumor cells to DNA-damaging agents.9–15 Because pan-HDAC inhibitors may induce also acetylation of p53,16, 17 thus influencing its function in cell DNA damage response, in this study, we used an ovarian carcinoma cell system including p53 mutant platinum-resistant sublines. The present results, indicating an accumulation of DNA damage following combined treatment with the atypical retinoid/HDAC inhibitor, provide additional evidence of the genotoxic stress as a primary mechanism by which atypical retinoids induce apoptosis in ovarian carcinoma cells.

Figure 1. Chemical structure of RC307.

Material and Methods

Drugs and antibodies

ST1926 and RC307 (N-hydroxy-3-(4′-hydroxybiphenyl-4-yl)-acrylamide, also known as ST2782) were prepared as described.18, 19 All the compounds were dissolved in DMSO and further diluted in culture medium (final concentration 0.5%).

The used primary antibodies were against: p53 (Dako, Glostrup, Denmark); phospho p53 (ser¹⁵), acetylated p53 (lys³⁸²), cleaved caspase 3 (asp¹⁷⁵), phospho Chk2 (thr⁶⁸), phospho Chk1 (ser345) (Cell Signalling Technology, Beverly, MA); RPA-2 and p21 (Neomarker, Union City, CA); PARP-1 (Oncogene Science, Uniondale, NY); γ-H2AX (Upstate Biotechnology, Lake Placid, NY); actin, β-tubulin and acetylated α-tubulin (Sigma, St., MO); cytochrome C (BD Pharmingen/Becton Dickinson); phospho-ser 1981 ATM (Lake Placid Biologicals, Lake Placid, NY); Chk1 (sc-8408, Santa Cruz Biotechnology, Santa Cruz, CA); Chk2 (provided by Dr. D. Delia, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy).

Cell lines and culture conditions

The human ovarian carcinoma cell lines IGROV-1, IGROV-1/Pt1 (cisplatin-resistan subline) and IGROV-1/OHP (oxaliplatin-resistant subline) were maintained in RPMI-1640 (Lonza, Verviers, Belgium) supplemented with 10% fetal calf serum (Invitrogen, Carlsbad, CA). The platinum-resistant sublines were selected by exposure to increasing cisplatin or oxaliplatin concentrations, respectively.20, 21

Antiproliferative activity

Cells were seeded in duplicate into 6-well plates. Twenty-four hours after seeding, cells were exposed to ST1926 and RC307, alone or in combination for 72 hr. After treatment, adherent cells were trypsinized and counted by a cell counter (Coulter Electronics, Luton, United Kingdom). Analysis of the interaction between drugs was performed by a modified method of Drewinko et al.22 Drewinko index >1 indicates a supra-additive effects. In all experiments, control cells were treated with the solvent at the same concentration used for drug-exposed cells.

Cell-cycle distribution and apoptosis

For cell-cycle analysis, cells were trypsinized, fixed in 70% ethanol, stained in phosphate-buffered solution (PBS) containing 10 μg/ml propidium iodide (Sigma) and RNase A (66 U/ml, Sigma) for 18 hr, analyzed by FACScan flow cytometer (Becton Dickinson, Mountaion View, CA).

Apoptosis was detected by transferase-mediated nick end labeling (TUNEL) assay and by annexin V-binding assay. For TUNEL assay, adherent and floating cells were harvested and analyzed for DNA fragmentation by terminal deoxynucleotidyl TUNEL assay (Roche, Mannheim, Germany) according to the manufacturer's recommendation. Apoptosis was assessed by flow cytometer, and the results were analyzed using the CellQuest software. In the annexin V-binding assay, cells were suspended in 100 μl of binding buffer (10 mM Hepes-NaOH, pH 7.4; 140 mM NaCl, 2.5 mM CaCl₂) and 5 μl of annexin (Becton Dickinson) and 5 μl of propidium iodide (2.5 μg/ml) for 15 min. Samples were analyzed by flow cytometry after the addition of 400 μl of binding buffer.

Western blot analysis

Adherent and floating cells were lysed in hot sodium dodecyl sulfate (SDS) sample buffer as described.5, 7 Whole-cell lysates (40 μg) were separated by SDS polyacrylamide gel electrophoresis and transferred onto nitrocellulose filters. Blots were preblocked for 2 hr at room temperature in PBS containing 5% (w/v) dried nonfat milk and incubated with primary antibodies overnight and with peroxidase-conjugated anti-mouse or anti-rabbit antibodies to reveal immunoreactive bands using the enhanced chemiluminescence detection system from Amersham Biosciences (Amersham, United Kingdom).

Measurement of cytochrome C release from mitochondria

Cells were washed in ice-cold PBS and processed as already described.7 Briefly, cells were lysed on ice cold lysis buffer (20mM HEPES-KOH, pH 7.5, 10 mM KCl, MgCl₂, 1 mM EDTA, 1 mM EGTA, 250 mM sucrose, 0.5 mM phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 10 μg/ml aprotinin and 10 μg/ml trypsin inhibitor) for 15 min in ice. Cells were homogenized using a glass Dounce and pestle for around 50 strokes and centrifuged at 20,000g for 20 min. Supernatant were stored at −20°C. Cytosolic protein extracts were analyzed by Western blot as described above.

Antitumor activity in vivo

All experiments were carried out using female athymic Swiss nude mice, 6–8-weeks-old (Charles River, Calco, Italy). Mice were maintained in laminar flow rooms keeping temperature and humidity constant. Mice had free access to food and water. Experiments were approved by the Ethics Committee for Animal Experimentation of the Fondazione IRCCS Istituto Nazionale dei Tumori of Milan according to institutional guidelines and to the UK Coordinating Committee on Cancer Research Guidelines.23

For the solid IGROV-1 tumor model, exponentially growing tumor cells (10⁷ cells/mouse) were s.c. injected into mice flank. Tumor lines were achieved by serial s.c. passages of fragments (about 2 × 2 × 6 mm) from growing tumors into nude mice. Groups of 5–8 mice bearing s.c. tumors implanted in 1 or both flanks were used. Tumor fragments were implanted on day 0, and tumor growth was followed by biweekly measurements of tumor diameters with a Vernier caliper. Tumor volume (TV) was calculated according to the formula: TV (mm³) = d² × D/2 where d and D are the shortest and the longest diameter, respectively. Drugs were delivered per os by gavage starting when tumors were just palpable. The efficacy of the drug treatment was assessed as: TV inhibition (TVI%) in treated versus control mice, calculated as: TVI% = 100 − (mean TV treated/mean TV control × 100). The toxicity of the drug treatment was determined as body weight loss and lethal toxicity. For statistical comparison of TVs in treated versus control mice the Student's t-test (2-sided) was used.

In the intraperitoneal ovarian carcinoma model, IGROV-1 ascitic cells, maintained by serial intraperitoneal passage, were injected i.p. in a volume of 0.2 ml containing 2.5 × 10⁶ cells. Tumor grows as ascites and small solid masses causing hemorrhagic and diffuse carcinomatosis. Drug efficacy was assessed in terms of delay of i.p. disease onset and rate of disease-free animal at the end of the experiment (45 days after tumor cell injection).

Results

Cell-growth inhibition studies

In this study, we chose IGROV-1 ovarian carcinoma cells as a model system based on the following considerations. IGROV-1 cells carrying wild-type p53 are susceptible to DNA damage-induced apoptosis.5 Platinum-resistant sublines used for comparison are characterized by loss of p53 function as a consequence of mutation20, 21 and somewhat less sensitive to ST1926 than the parental cells.5 In combination studies, we used RC307, a novel HDAC inhibitor with a hydroxamic acid-based structure characterized by a broad-spectrum inhibitory activity.19 The parental IGROV-1 line and platinum-resistant sublines exhibited a comparable sensitivity to RC307 with IC50 values in the range of 6.5–7.5 μM. Using a subtoxic (3 μM; i.e., ∼IC₂₀) or toxic (10 μM, ∼IC₇₀) concentration of RC307, Western blot analysis revealed a dose-dependent acetylation of histone H4, α-tubulin and p53, proteins known to be substrates of various HDAC isoforms16 (Fig. 2). These effects were already evident after 4 hr-exposure and did not increase remarkably after 24 hr. The accumulation of acetylated proteins was less marked in IGROV-1/Pt1. RC307 also produced a downregulation of p53, which was almost complete at 24 hr in Pt-resistant cells treated with 10 μM. This effect was consistent with the expression of a mutant p53 in resistant cells and suggested destabilization of the mutant protein likely as a consequence of HSP90 acetylation.

Figure 2. Protein acetylation in ovarian carcinoma cells treated with RC307. Acetylation of p53, H4 and α-tubulin was analyzed by Western blot analysis after 4 or 24 hr of treatment. β-Tubulin is shown as a control of protein loading. The results of 1 representative of 2 experiments are shown.

The combination treatment of RC307 with ST1926 (used at 2 dose levels, IC₃₀ and IC₇₀) resulted in a stronger cell-growth inhibition when compared with single agent (Fig. 3). Indeed, a subtoxic (∼IC₃₀) dose of ST1926 enhanced cell-growth inhibition in a wide range of doses of RC307. The sensitization was still evident at low concentrations of RC307 (i.e. ≤1 μM), which alone caused negligible effects. The analysis of drug interaction supported a synergistic interaction between ST1926 and RC307 (Table 1). The potentiation of ST1926 effect was less evident in IGROV-1/Pt1.

Figure 3. Dose-response curves for the antiproliferative effects of RC307 in IGROV-1, IGROV-1/OHP and IGROV-1/Pt1 cells in the absence (▪) or presence of ST1926 at concentrations corresponding to IC₇₀ (▴, 0.3 μM in IGROV-1, 0.4 μM in IGROV-1/OHP and 0.6 μM in IGROV-1/Pt1) or to IC₃₀ ([▾], 0.15 μM in IGROV-1, 0.2 μM in IGROV-1/OHP and 0.3 μM in IGROV-1/Pt1). The growth-inhibitory effect was determined after 72 hr of treatment. The results are the mean of 3 independent experiments ± SD.

Table 1. Analysis of drug interaction for the combined treatment of ST1926 and RC370 (72 hr) in ovarian carcinoma cell lines

Cell-cycle perturbation

At concentrations, which produced predominantly antiproliferative effects, the treatment with ST1926 or RC307 alone produced a marginal perturbation of cell-cycle progression with a moderate increase of S-phase cells (Fig. 4a). A simultaneous exposure to the 2 drugs determined a persistent G1 phase arrest and, appearance of sub-G1 phase, thus suggesting the induction of apoptotic cell death. The effects of the combination of ST1926 and RC307 (10 μM) on the cell-cycle distribution were similar to that induced by highly cytotoxic concentrations of ST1926 itself after 24 hr of treatment (Fig. 4b).

Figure 4. Cell-cycle distribution of IGROV-1 cells and platinum-resistant sublines. Cells were treated with (a) ST1926 (IC₃₀: 0.15 μM in IGROV-1, 0.2 μM in IGROV-1/OHP and 0.3 μM in IGROV-1/Pt1) or RC307 (3 and 10 μM), alone or in combination; (b) higher cytotoxic concentrations (IC₇₀) of ST1926 alone. Cell cycle was assessed by fluorescence-activated cell sorting analysis of propidium iodide-stained cells. Representative profiles of at least 3 independent experiments are shown.

Sensitization to apoptosis by ST1926/RC307 combination

Under the same treatment conditions used for analysis of cell-cycle perturbation, exposure to each single agent resulted in a barely detectable caspase 3 activation (Fig. 5), thus indicating that, at the tested concentrations, the effect of single-drug treatment was predominantly cytostatic. In contrast, cells treated with the drug combination exhibited a dose-dependent activation of caspase 3, associated with cleavage of PARP-1, which was already detectable after 24 hr. Moreover, consistent with an enhanced activation of caspase-dependent apoptotic pathways was the observation that the combination of ST1926 with RC307 caused the cytosolic release of cytochrome C, a caspase activator (Fig. 5b). The time-course of apoptosis induction, determined by TUNEL assay (Table 2), revealed a marginal (if any) induction of apoptosis in the presence of each single agent. On the contrary, the combined treatment produced a remarkable increase of apoptosis by 48 hr in all cell lines. This pattern of apoptosis response was confirmed by the Annexin V-binding assay (Fig. 6). The onset of apoptosis was more rapid in IGROV-1 cells characterized bya functional wild-type p53, because apoptosis was already detectable at 24 hr.

Figure 5. Analysis of apoptosis-related factors. Cells were exposed to ST1926 (0.15 μM in IGROV-1, 0.2 μM in IGROV-1/OHP and 0.3 μM in IGROV-1/Pt1) and to RC307 (3 and 10 μM), alone or in combination for the indicated times. (a) Cleavage of caspase 3 (CPP32) and PARP-1. Whole-cell lysates were analyzed by Western blotting with the indicated specific antibodies; β-tubulin is shown as control for protein loading. (b) Release of cytochrome C into the cytosol. The cytosolic extracts were prepared after the indicated times of drug exposure. One representative of 2 experiments is shown. Control (1), cells treated withST1926 (2) or RC307 3 μM (3) or RC307 10 μM (4) or ST1926 + RC307 3μM (5) or ST1926 + RC307 10 μM (6).

Figure 6. Induction of apoptosis in IGROV-1 cells and platinum-resistant sublines treated with ST1926 and RC307 alone or in combination, for 72 hr. ST1926 was used at equitoxic concentration, that is, IC₃₀ (0.15 μM in IGROV-1, 0.2 μM in IGROV-1/OHP and 0.3 μM in IGROV-1/Pt1). Apoptosis was measured by flow cytometric analysis of annexinV binding. Cells with early apoptotic manifestations are in the low-right quadrant (Annexin V-positive, propidium iodide-negative), whereas necrotic cells and cells with late apoptotic manifestations are in the up-right quadrant (Annexin V-positive, propidium iodide-positive). One representative experiment of 2 is shown.

Table 2. Apoptosis induction in IGROV-1, IGROV-1/PT1 and IGROV-1/OHP cells by RC307 or ST1926 alone, or in combination (tunel assay)

DNA damage response

Because ATM is a key player in DNA damage response,24 we examined its activation in IGROV-1 cells treated with antiproliferative or cytotoxic concentrations of ST1926 (0.15 μM, i.e., IC₃₀ and 0.3 μM, i.e. IC₇₀) by immunofluorescence analysis. Foci of phosphorylated ATM (Ser1981) were found following 4-hr exposure (Fig. 7a). The analysis of the fluorescence intensity by FACscan documented a dose-dependent increase of ATM phosphorylation. This finding was consistent with previous observations indicating the drug-induced formation of DNA double-strand breaks.5, 7 A similar effect was observed in resistant cells.

Figure 7. Activation of ATM, Chk1 and Chk2 in response to ST1926. (a) Immunofluorescence staining for phospho-ATM (ser1981). Four hours after ST1926 treatment (IC₃₀ and IC₇₀), cells were stained with mouse anti-pospho-ATM (ser1981) antibodies and AlexaFluor-conjugated goat antimouse 488 Ig second antibodies (green). Nuclei were stained with Hoechst 33342 (blue). Bar, 5 μm. Cells were also analyzed by flow cytometry, and the percent values are referred to the phospho-ATM (ser1981) positive cells. Results of 1 representative experiment of 2 are shown. (b) Western-blot analysis performed at 4 or 24 hr after ST1926 treatment. β-Tubulin is shown as a control of loading. The results of 1 of 2 representative experiments are shown.

Because ATM phosphorylates Chk1 and Chk2,22, 24 which are multifunctional enzymes involved in the induction of cell-cycle arrest and apoptosis by DNA damage,25, 26 the phosphorylation of these proteins was investigated by Western blot analysis in IGROV-1 cells (Fig. 7b). The phosphorylation of Chk1 was an early dose-dependent event, detectable after 4-hr exposure. The phosphorylation of Chk2 was less marked but more persistent, still detectable at 24 hr. This event was concomitant with a partial downregulation of Cdc25A, a phosphatase substrate of Chk1 and Chk2, which undergoes degradation following phosphorylation.

On the basis of the effects of ST1926 on proteins implicated in DNA damage checkpoint control, we examined the cellular response to the combination of ST1926 and RC307 (Fig. 8). No detectable phosphorylation of ATM was observed in untreated and RC307-treated cells, whereas ST1926 induced an intense nuclear clustering of phosphorylated ATM at serine 1981 and the combination with RC307 caused an enhanced and persistent phosphorylation of ATM. The combination after 4-hr exposure induced an increase of the phosphorylation of Chk1 at ser345 as compared with the single drugs (Fig. 9). Chk2 phosphorylation was detectable only later (24 hr).

Figure 8. Immunofluorescence staining for phospho-ATM (ser1981). IGROV-1 cells were exposed to the ST1926 IC₃₀ (0.15 μM) and RC307 (10 μM), alone or in combination for 4 hr. Cells were stained with mouse anti-pospho-ATM (ser1981) antibodies and AlexaFluor-conjugated goat anti-mouse 488 Ig second antibodies (green). Nuclei were stained with Hoechst 33342 (blue). Bar, 5 μm. Upper panel shows a representative staining for control and treated cells. Lower panel reports the percentage of phospho-ATM positive cells in treated and untreated cells. The results, expressed as a percentage of the total cell population, are the mean of at least 2 independent experiments ± SD.

Figure 9. Effects of ST1926, RC307, alone or in combination, on activation of DNA damage checkpoint proteins in IGROV-1 cells and in platinum-resistant sublines. Cells were exposed to ST1926 (0.15 μM in IGROV-1, 0.2 μM in IGROV-1/OHP and 0.3 μM in IGROV-1/Pt1) and to RC307 (3 and 10 μM), alone or in combination. Control (1), cells treated with ST1926 (2) or RC307 3 μM (3) or RC307 10 μM (4) or ST1926 + RC307 3 μM (5) or ST1926 + RC307 10 μM (6). β-Tubulin is shown as a control of loading. One representative of 2 experiments is shown.

A critical substrate of ATM is histone H2AX, which is rapidly phosphorylated in response to DNA double-strand breaks and its phosphorylation is a sensitive marker of DNA damage.27 Indeed, following exposure to ST1926 (IC₃₀), an early phosphorylation of histone H2AX was detected by Western blot analysis in IGROV-1 cells and sublines (Fig. 9). The drug combination resulted in an enhanced phosphorylation, which was most evident at 24 and 48 hr of exposure in IGROV-1 and IGROV-1/OHP. H2AX phosphorylation was less persistent in IGROV-1/Pt1 cells, suggesting an increased DNA repair efficiency.

As an additional marker of DNA damage response, we determined the phosphorylation state of RPA-2, a DNA single-strand breaks binding protein, which is rapidly phosphorylated after DNA damage (Fig. 9). A slower migrating band, evidenced by the shift in the electrophoretic mobility and corresponding to phosphorylated RPA-2, was barely detectable in ST1926-treated cells at 4 hr but more evident in cells treated with the combination. A prolonged RPA-2 phosphorylation, still detectable after 48-hr exposure (not shown), was found in cells treated with the combination.

p53 phosphorylation in serine 15 and acetylation in lysine 382 were assessed as additional indicators of DNA damage response in IGROV-1 expressing wild-type p53. The state of phosphorylation and acetylation of p53 is known to be crucial in triggering the apoptotic response following genotoxic stress.17 As shown in Figure 10, ST1926 induced p53 phosphorylation at serine 15 and acetylation at lysine 382 after 4 hr of treatment. These effects increased in the combined treatment and were persistent up to 48 hr.

Figure 10. Modulation of p53. Expression of phospho-p53 (ser15) and acetyl-p53 (lys382) was analyzed in IGROV-1 cells treated with ST1926 (0.15 μM) or RC307 (3 or 10 μM), alone or in combination. Whole cell extracts were prepared and analyzed by Western blot analysis. β-Tubulin is shown as a control of loading. One representative of 2 experiments is shown. Control (1), cells treated with ST1926 (2) or RC307 3 μM (3) or RC307 10 μM (4) or ST1926 + RC307 3 μM (5) or ST1926 + RC307 10 μM (6).

In vivo antitumor activity

The antitumor efficacy of ST1926 and RC307 combination was examined in the IGROV-1 tumor growing in athymic nude mice as s.c. implanted solid tumor or as orthotopic model following i.p. tumor cell injection (Table 3). In the treatment of the solid tumor, using oral doses of both agents, which alone produced a marginal growth inhibition, only the combination resulted in a significant antitumor effect (p < 0.05). In the treatment of the ovarian carcinoma implanted i.p., the oral administration of RC307 did not produced appreciable antitumor effects. ST1926 produced a marked delay in disease progression, and the combination of RC307 with ST1926 resulted in a substantial improvement of efficacy at least in term of delay of manifestations of i.p. disease. The effect of the in vivo combination was not impressive, likely as a consequence of the low bioavailability of oral RC307. Further in vivo evaluation of the efficacy of the combination requires optimization of the formulations of RC307. Preliminary observations indicated that the methoxy derivative of RC307 was characterized by an improved metabolic stability and pharmacological behavior and could provide a better therapeutic outcome.19

Table 3. Antitumor activity of ST1926, RC307 alone or in combination against the ovarian carcinoma model IGROV-1

Discussion

The therapeutic potential of the atypical retinoid ST1926 in the treatment of ovarian carcinoma has been ascribed to its proapoptotic activity.3 Indeed, ST1926 was found to be a potent inducer of apoptosis in ovarian carcinoma cells.5 The primary intracellular target(s) of ST1926 remains to be identified and the molecular events implicated in apoptosis induction are not clearly defined. The results reported in this study provide additional support to our previous observations supporting genotoxic stress as a critical event implicated in ST1926-mediated cell death.5, 7 Indeed, several lines of evidence are consistent with the interpretation that apoptosis occurred downstream of DNA-damaging events. ST1926 induces early (2–4 hr) increase in γ-H2AX foci in various cell lines including IGROV-1 ovarian carcinoma cells used in this study. The pattern of cellular response to drug treatment is reminiscent of a typical DNA damage response.

The results of this study indicated that phosphorylation of H2AX, one of the first events in response to induction of double-strand breaks,24, 26, 27 occurred concomitant with activation of ATM, a key regulator of the signaling network of the cellular response to double-strand DNA lesions. These events are not the consequence of a rapid apoptosis triggered by the drug, because, in contrast to leukemia cells, in our cell systems, biochemical manifestations of apoptosis (e.g., cytochrome C release and caspase activation) could be barely detected at 24 hr. In accord with ATM activation, ST1926 induced a concomitant (dose-dependent) phosphorylation of Chk1 and Chk2 (Fig. 7), which are known to mediate the checkpoint functions of ATM in response to agents inducing DNA double-strand breaks (e.g. ionizing radiation).24 Again, consistent with the activation of ATM pathway is the phosphorylation of p53 (Fig. 10).

In an attempt to gain further insights into the mechanism by which atypical adamantyl retinoids and, more specifically, ST1926, exhert proapoptotic effects, in this study, we have explored the ability of a pan-HDAC inhibitor to sensitize ovarian carcinoma cells to ST1926, because HDAC inhibitors have been reported to enhance DNA damage induced by various cytotoxic agents including ionizing radiation.9–15 Indeed, because histone acetylation modulates chromatin structure and gene expression, HDAC inhibition may influence DNA damage response and ATM-mediated DNA damage response depends on chromatin organization.28 However, the mechanism(s) by which HDAC inhibitors enhance the cytotoxicity of genotoxic agents is not well understood, because the acetylation of nonhistone proteins may also be implicated in the pleiotropic effects of HDAC inhibitors.16 The present findings provide evidence that the combination of ST1926 and the HDAC inhibitor RC307, at concentrations which produced only cell-growth inhibition, resulted in a marked sensitization of ovarian carcinoma cells to apoptosis (Table 2). The synergistic interaction between ST1926 and RC307 was also supported by the enhanced release of cytochrome C from mitochondria (Fig. 5b) and by a substantial activation of caspase 3 and cleavage of PARP-1 (Fig. 5a). The sensitization was associated with a dose-dependent increase of phosphorylation of ATM, Chk1 and a delayed phosphorylation of Chk2. The most evident effect of the combination was the increase and prolongation of histone H2AX phosphorylation, which is known to form nuclear foci at the sites of DNA double-strand breaks.26 Treatment with ST1926, but not with RC307, induced a rapid phosphorylation of H2AX. The combination with the HDAC inhibitor caused a marked increase of the γH2AX phosphorylation at 24 hr. A similar response has been described in the combination of vorinostat, an HDAC inhibitor with a profile of activity comparable to RC307, with ionizing radiation and has been interpreted as an inhibition of the repair process.10 The persistence of the ST1926-induced DNA lesions was also consistent with an increased and persistent phosphorylation and acetylation of p53 detectable up to 48 hr (Fig. 10). The presence of a functional wild-type p53 did not appear to be required for activation of apoptosis, because also ovarian carcinoma cells carrying mutant p53 were able to activate apoptosis. However, the observation that the presence of functional p53 conferred a susceptibility to a more rapid onset of apoptosis, as indicated by the response of IGROV-1 cells (Table 2), supports that the activation of p53 provides a contribution in sensitization for apoptosis of wild-type p53 cells. The somewhat heterogeneous response of the p53 mutant Pt-resistant sublines likely reflected various additional alterations generated during selection with different Pt drugs. For example, IGROV-1/Pt1 exhibits an increased DNA damage tolerance.21

In conclusion, although the mechanism(s) by which the adamantyl retinoid ST1926 cause DNA damage remains unclear, the present findings, indicating an enhanced DNA damage response by the combination with RC307, strongly support the role of genotoxic events in the mechanism of apoptotic activity of this agent. A plausible explanation of the sensitization by the HDAC inhibitor could be its ability to modulate the chromatin structure resulting in increased exposure to DNA-damaging agents. Thus, the observation indicating sensitization to apoptosis and enhanced DNA damage response following the combination of the atypical retinoid and the HDAC inhibitor provides additional support to our interpretation implicating DNA damage in the mechanism of action of adamantyl retinoids.5–7 Finally, because the sensitization of ovarian carcinoma cells to ST1926 by the HDAC inhibitor was observed in a wide range of doses, including subtoxic doses, these findings may have relevance in the design of rational combinations for the clinical use of ST1926. The therapeutic interest of this combination is supported by preliminary in vivo experiments.