Sensitivity of osteosarcoma cells to HDAC inhibitor AR-42 mediated apoptosis

The HDAC inhibitors AR-42 and SAHA were synthesized as described [10]. Stock solutions AR-42 and SAHA were prepared in DMSO and diluted in the indicated culture medium for treatment of cells in vitro. Antibodies against Akt, pAkt-Ser473, phosphor-glycogen synthase kinase-3 (GSK3)β, caspase-3, Bcl-xl, α-tubulin, cyclin D1, phosphor-p70 ribosomal protein S6 kinase (p70S6K), p-mTOR and PTEN were purchased from Cell Signaling Technologies (Beverly, MA). Additional polyclonal rabbit antibodies used were acetylated histones H3 (N-terminus) and H4 (Lys5/8/12/16) (Upstate Biotechnology, Inc., Lake Placid, NY), β-actin (Sigma, St. Louis, MO), and survivin (Novus Biologicals, Littleton, CO).

Canine OS cell lines, OSCA-2, −7.2, −16, −36, −39.1, −40, −50, were previously established in the laboratory of one of the authors (JFM). Normal canine osteoblasts were obtained from Cell Applications, Inc. (San Diego, CA). The canine OS cell line D17 and human OS cell lines SAOS-2, SJSA and U2OS were obtained from American Type Culture Collection (Manassas, VA). Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) (OSCA-2, −16, −36, −39.1, −40, −50), RPMI-1640 (D17, OSCA-7.2), or McCoy’s medium (Gibco, Invitrogen, Carlsbad, CA) (SJSA, SAOS-2 and U2OS) supplemented with 10% fetal bovine serum (FBS, HyClone, Gemini, West Sacramento, CA) and antibiotics (100 U/ml penicillin, 0.1 mg/ml streptomycin) (Gibco) in a humidified incubator containing 5% CO₂ at 37 °C.

The effect of AR-42 and SAHA on the viability of canine and human tumor cells was assessed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Sigma) as described previously [18]. Briefly, cells were seeded in 96-well plates at ~2500 cells per well in medium supplemented with 10% FBS. After 24 h, the cells were treated with varying concentrations (1–10 μM) of AR-42 and SAHA, dissolved in DMSO (final DMSO concentration ≤ 0.1%) for 24, 48 and 72 h. Controls were treated with DMSO vehicle alone at a concentration equal to that of drug-treated cells. After drug treatment, 22 μl of MTT reagent (5 mg/ml) was added to each well and the cells were incubated for up to 2 to 4 h at 37 °C. The absorbance was read on a plate reader (UV Spectromax M2 plate reader, Molecular Devices, Sunnyvale, CA) at 570 nm. The concentration of AR-42 and SAHA that inhibited cell viability by 50% (IC₅₀) was determined using CompuSyn software (v. 3.0.1, ComboSyn, Inc., Paramus, NJ) and the values expressed as mean ± SD. All treatments were evaluated in triplicate in at least three independent experiments.

Cells were exposed to AR-42 or SAHA at 1 and 10 μM concentrations for 48 h, washed with phosphate-buffered saline (PBS), resuspended in 500 μl of cold PBS, and then added drop wise to 70% ethanol and stored at 4 °C overnight. Cells were then washed twice in PBS and suspended in 500 μl of PBS containing 10 ug/ml of RNase A (LC, Laboratories, Woburn, CA) and 50 ug/ml of propidium iodide (Sigma) and assessed by BD FACS Calibur (Becton-Dickinson, San Jose, CA). Data were analyzed by Cell Quest flow software (Becton-Dickinson). A maximum of 10,000 cells within the gated region were analyzed for each treated and untreated sample. Experiments were replicated three times.

Drug-induced apoptotic cell death was determined by detection of DNA fragmentation using the Cell Death Detection ELISA kit (Roche, Indianapolis, IN). The ELISA was performed according to the manufacturer’s protocol and is based on the quantitative determination of cytoplasmic histone-associated DNA fragments in the form of mononucleosomes or oligonucleosomes generated after induced apoptotic death. Briefly, 3 × 10⁵ cells were cultured in RPMI/McCoy’s medium supplemented with 10% FBS in 100 mm tissue culture dishes for 24 h before treatment. Cells were treated with varying concentrations of AR-42 and SAHA (1–10 μM) and DMSO vehicle as control for 48 h. Approximately 100,000 cells were used per assay. The absorbance was read on a plate reader (UV Spectromax M2 plate reader, Molecular Devices) at a wavelength of 405–490 nm.

Activation of the caspase 3/7 pathway following the drug treatment was measured by Sensolyte™ Homogeneous AMC Caspase-3/7 Assay kit (AnaSpec, San Jose, CA) following the manufacturer’s protocol. Briefly, 1x10⁵ cells per well were seeded in six-well plates and treated for 48 h at concentrations of 1 and 10 μM AR-42 or SAHA. Each treatment was performed in triplicate. After treatment for the indicated time, cells were lysed with lysis buffer to a final volume of 150 μl/well and 50 μl of caspase 3/7 substrate reagent was then added to the wells and incubated on a plate shaker for 30–60 s at 100–200 rpm. Fluorescence were read on a plate reader (UV Spectromax M2 plate reader, Molecular Devices) at an Ex/Em = 354 nm/442 nm and recorded after 1 h.

For immunoblotting analysis, drug-treated (AR-42 and SAHA at 1 and 10 μM) and vehicle (DMSO)-treated cells were collected 48 h after treatment, washed in PBS, and lysed in M-PER protein extraction reagent (Pierce Biotechnology, Rockford, IL) unless otherwise stated. After centrifugation at 14,000 rpm for 15 min equal amounts of total protein from the cell lysates were resolved on 4–20% denaturing polyacrylamide gels (Invitrogen) and transferred to nitrocellulose membranes (PALL-Germany). After blocking with TBST containing 5% non-fat dry milk (Blotto, BioRad Laboratories, Hercules, CA) for 1 h, the membranes were incubated with indicated primary antibodies at 4 °C overnight and then washed three times with TBST. The membranes were probed with horseradish peroxidase conjugated secondary antibodies (Jackson Immune Research Laboratories, West Grove, PA) for 1 h at room temperature and washed three times with TBST. The blots were then developed with Western Lightning reagents (Perkin-Elmer, Waltham, MA).

The effect of combining AR-42 with doxorubicin on cancer cell viability was evaluated in U2OS and D17 cells using the fixed-ratio method. Cells were treated with AR-42 and doxorubicin individually and in combination. For combination treatment of U2OS cells, drugs were combined at a concentration ratio of 3.3:1 (AR42:doxorubicin) and 2-fold serial dilutions were performed to generate a series of solutions containing AR-42 and doxorubicin at concentrations ranging from 0.125 to 8 μM and 0.0375 to 2.4 μM, respectively. For treatment of D17 cells, the concentrations of AR-42 in the combination ranged from 0.625 to 40 μM, and that of doxorubicin ranged from 0.04125 to 2.64 μM to yield a fixed concentration ratio of 15.2:1 (AR-42:doxorubicin). After treatments, cell viability was determined by MTT assays as described above. Dose-effect data for individual drugs and their combinations were analyzed for synergistic effects using the median-effect method of Chou and Talalay [19] using CompuSyn software (v. 3.0.1, ComboSyn, Inc.). Combination index (CI) values were calculated to characterize the nature of the drug interaction as defined by Chou and Talalay: CI = 1, additivity; CI < 1, synergism; CI > 1, antagonism. The dose reduction index (DRI) is a measure of the extent to which the dose of a drug in a synergistic combination is reduced, compared with the dose of the same drug alone, to achieve a given effect level. The DRI value for each drug was also calculated.

Cell viability was measured in triplicate and averaged to achieve a single value for each combination of drug treatment, concentration, and cell line. The results were plotted over the three measurement periods and visually inspected (Fig. 1c). The cell viability at 72 h was compared by drug treatment assignment, AR-42 (n = 24) versus SAHA (n = 24). Samples were tested for normality (D ’Agostino-Pearson test) and equality of variance (Levene’s test). Inasmuch as samples were not normally distributed, a Mann-Whitney test for independent samples was performed to compare cell viability by drug treatment group. Analyses were performed using commercial software (MedCalc Statistical Software version 16.8, MedCalc Software bvba, Ostend, Belgium). Statistical significance was set at P < 0.05.

The antiproliferative effects of AR-42 and SAHA were assessed in three human (SAOS-2, SJSA and U2OS) and eight canine (D17, OSCA-2, −7.2, −16, −36, −39.1, −40 and −50) OS cell lines. OS cells were treated with each drug over a concentration range of 0.1 to 10 μM and cell viability was measured by MTT assay at 24, 48 and 72 h of treatment. AR-42 caused dose-dependent reductions in cell viability in all of the human cell lines starting at 48 h of treatment, except in SJSA cells for which a decrease in viability was not seen until 72 h (Fig. 1a; 24 and 48 h data for SJSA are not shown). Similarly, AR-42 reduced cell viability in most of the canine cell lines, which, however, exhibited differential sensitivities ranging from near complete resistance (OSCA-39.1) to IC50 values of <1 μM at 72 h of treatment (D17, OSCA-2 and OSCA-50) (Fig. 1b). OSCA-16, −7.2 and −36 cell lines were relatively resistant. The IC50 values for sensitive human and canine OS cells ranged from 0.25 to 2.2 μM after 72 h of treatment. Comparison of the antiproliferative activities of AR-42 and SAHA was performed in sensitive human (U2OS, SAOS-2) and canine (D17, OSCA-2) OS cell lines which showed that both inhibitors induced time- and dose-dependent decreases in cell viability and that, in all the cell lines tested, AR-42 was more potent than SAHA (Fig. 1c). Finally, the sensitivity of non-malignant canine osteoblasts to AR-42 relative to human (SAOS-2, U2OS) and canine (OSCA-2, D17) OS cells was determined. As shown in Fig. 1d, the median percent cell viability at 72 h was significantly greater for SAHA-treated cells (Z = 2.474, P = 0.0133). This effect appeared to be present at all dilutions of the drugs and in all cell-lines, although the magnitude of the differences varied with combinations of drug concentration and cell lines.

The antitumor effects of HDAC inhibitors have been linked to their ability to inhibit growth and induce apoptosis in cancer cells. To determine the ability of AR-42 to induce apoptosis in OS cells, we analyzed the effects of AR-42 and SAHA treatment on OS cells by cell cycle analysis, quantitation of nucleosome fragmentation, and measurement of caspase activity. Both inhibitors induced dose-dependent apoptosis as indicated by an increased sub-G1 cell population after 48 h of drug treatment, which inversely correlated with a reduced percentage of cycling cells, confirming decreased cell proliferation (Fig. 2a). Relative to SAHA, AR-42-treated cells exhibited a significantly greater increase in apoptosis at the 1 μM concentration in all cell lines tested (Fig. 2a). Increases in nucleosome fragmentation paralleled those in the subG1 cell population (Fig. 2b). Furthermore, both HDAC inhibitors increased caspase 3/7 enzymatic activity in all cell lines, consistent with activation of the intrinsic apoptotic pathway (Fig. 2c). Lastly, Western blotting demonstrated a dose-dependent increase in caspase 3 cleavage and a decrease in Bcl-xl expression following HDAC inhibition (Fig. 2d). These results demonstrate that AR-42 induces apoptosis in both canine and human OS cell lines at low micromolar drug concentrations.

The antitumor effects of HDAC inhibitors are often the consequence of altered gene transcription resulting from increased histone acetylation. To examine the relationship between HDAC inhibition and histone acetylation in OS cells we assessed the levels of acetylated histones in OS cells following treatment with AR-42 or SAHA. Western blot analysis after 48 h of treatment revealed a dose-dependent induction of acetylated histones H3 and H4 in both human and canine OS cell lines (Fig. 3) demonstrating that these drugs actively modulate chromatin by inhibiting HDAC activity in the tumor cell lines. These effects on histone acetylation correlated with the differential and dose-dependent effects of AR-42 and SAHA on cell viability and apoptosis.

Both AR-42 and, to a lesser extent, SAHA have been shown to decrease Akt phosphorylation in prostate cancer cells [20] attributable to the disruption of HDAC-protein phosphatase-1 (PP1) complexes leading to increased PP1-Akt association [21]. Thus, we investigated the effects of both drugs on Akt phosphorylation and downstream markers of Akt signaling. Western blot analysis revealed that, after 48 h of treatment with 10 μM of either AR-42 or SAHA, pAkt-Ser⁴⁷³ was reduced in 2 of 3 human and 2 of 4 canine OS cell lines (data for D17 and SAOS-2 cells are shown, Fig. 4). Similarly, decreases in the phosphorylation of GSK3β, a direct phosphorylation target of Akt, were also observed in cells treated with higher concentrations of drug, which confirmed the functional suppression of Akt signaling in these HDAC inhibitor-treated cells. HDAC inhibitor treatment also caused reductions in the phosphorylation of mTOR and its substrate, p70S6K, more prominently in D17 cells than in SAOS-2 cells, in which the effects were small. The mTOR pathway is another downstream effector of Akt signaling and its dysregulation has been implicated in OS. At 1 μM AR-42 and SAHA, effects on Akt signaling were inconsistent, ranging from minor reductions to apparent lack of effect or even increases in these signaling markers, which was unexpected in light of clear effects of this concentration of drug on cell viability, apoptosis, and/or histone acetylation. Phosphatase and tensin homolog (PTEN) negatively regulates Akt signaling by antagonizing the action of phosphatidylinositol 3–kinase (PI3K) [22]. Thus, the effect of AR-42 on the expression of PTEN was also assessed. Neither AR-42 nor SAHA altered the abundance of PTEN protein in OS cell lines, suggesting that PTEN was not involved in the HDAC inhibitor-mediated suppression of pAkt-Ser⁴⁷³.

The up-regulation of survivin, an inhibitor of apoptosis protein (IAP) family member, has been reported to be associated with poor prognosis and increased risk of metastasis in OS [23‐25]. Previous studies have demonstrated that HDAC inhibitors, including AR-42, can suppress survivin expression in cancer cells [13, 20, 26]. Moreover, survivin expression has been shown to be regulated by the Akt/p70S6K1 pathway [27]. As shown in Fig. 4, treatment with either HDAC inhibitor resulted in a dose-dependent reduction of survivin in both human and canine OS cells. These data suggest that AR-42 induces apoptosis of OS tumor cell lines through modulation of key cellular proteins.

Several preclinical studies have described synergistic effects of HDAC inhibitors in combination with a diverse range of cytotoxic and targeted therapeutic agents [5]. AR-42 itself has been reported to sensitize prostate cancer and hepatocellular carcinoma cells to DNA double-strand break-inducing chemotherapeutic agents [28] and radiotherapy [29], respectively. Consequently, we investigated the potential synergistic activity of the combination of AR-42 and doxorubicin in human and canine OS cells. U2OS and D17 cells were treated with AR-42 and doxorubicin over a range of doses both as single agents and in combinations containing the drugs at fixed concentration ratios. The dose-effect data generated after single agent and combination treatments of each cell line (Fig. 5) were subjected to median-effects analysis from which CI (combination index) values were derived to define the drug interactions. As shown in Table 1, for both cell lines the AR-42/doxorubicin combination was synergistic, as all CI values were less than 1 at dose levels that reduced cell viability by 50% (effective dose-50, ED50) or greater. Calculation of the dose reduction index (DRI) values revealed that, as a result of the synergism between AR-42 and doxorubicin, their IC50 values could be reduced by 2.06-fold and 4.03-fold, respectively, in U2OS cells, and 1.40-fold and 4.57-fold, respectively, in D17 cells (Table 2).