Histone Deacetylase Inhibitors and 15-Deoxy-Δ^12,14-Prostaglandin J₂ Synergistically Induce Apoptosis

Histone deacetylase (HDAC) inhibitors (HDI) are promising anticancer agents that are potent inducers of growth arrest, differentiation, and apoptotic cell death in various types of tumor cell lines (1, 2). HDIs achieve these effects through the alteration of gene expression by inhibition of HDAC and subsequent changes in chromatin structure (3). Valproic acid (VPA) and suberoylanilide hydroxamic acid (SAHA, vorinostat) are HDIs and exhibit antitumor effects in vitro and in vivo (4–7). VPA, an anticonvulsant clinically used for decades, is currently under investigation in clinical studies for myelodysplastic syndromes (8). SAHA was approved in the United States for the treatment of cutaneous T-cell lymphoma in 2006. However, a considerable proportion of patients with cutaneous T-cell lymphoma did not respond well to SAHA alone, although ∼30% of patients overall achieved an objective response (9, 10). Recent clinical studies have shown that the combinations of VPA and all-trans retinoic acid, 5-aza-2′-deoxycytidine (decitabine), or the topoisomerase II inhibitor epirubicin are active in patients with hematopoietic malignancy or solid tumors (8, 11, 12). These combination therapies have led to exciting clinical responses in an otherwise refractory population. Therefore, additional mechanism-based therapeutic strategies for the use of HDIs in combination with other agents should be investigated to overcome resistance to HDIs or to increase the efficacy of HDIs.

Cyclopentenone prostaglandins (PG) are derived from arachidonic acid and are released from membrane phospholipids in response to various stress stimuli. It has been shown that the PGJ₂ series exert antitumor activity in vivo as well as in vitro (13, 14). PGD₂, which is the most abundant PG in normal tissue, is sequentially converted to PGJ₂ and then to Δ¹²-PGJ₂ or 15-deoxy-Δ^12,14-PGJ₂ (15d-PGJ₂) in vitro (15). In particular, 15d-PGJ₂ proved to be a natural endogenous ligand for peroxisome proliferator–activated receptor γ (PPARγ; refs. 16, 17) and the most potent antitumor agent of this class of PGs (18, 19). Moreover, 15d-PGJ₂ induces caspase-dependent and PPARγ-independent apoptosis through reactive oxygen species (ROS) generation (19, 20).

Here, we show that the cotreatments with HDIs and 15d-PGJ₂ synergistically induce apoptosis through the alteration of apoptosis-related genes through HDAC inhibition leading to ROS generation and subsequent endoplasmic reticulum (ER) stress. Moreover, the combined treatment with VPA and 15d-PGJ₂ significantly inhibits tumor growth in vivo as well. These results suggest that the combination of clinically relevant HDIs and the natural endogenous PPARγ ligand 15d-PGJ₂ will be promising for the treatment of a broad spectrum of malignant tumors in a clinical environment.

Human colon cancer DLD-1 cells and human chronic myeloid leukemia K562 cells were maintained in RPMI 1640. Human immortalized keratinocyte HaCaT cells and human colon cancer HT29 cells were maintained in DMEM. Culture media were supplemented with 10% fetal bovine serum, l-glutamine (2 mmol/L for RPMI 1640 and 4 mmol/L for DMEM), 100 U/mL penicillin, and 100 μg/mL streptomycin. Cell cultures were incubated at 37°C in a humidified atmosphere of 5% CO₂.

15d-PGJ₂, SAHA, and VPA were purchased from Cayman Chemical, Biomol, and Wako, respectively. N-acetylatedcysteine (NAC) and buthionine sulfoximine were purchased from Nacalai Tesque, Inc., and Wako, respectively. zVAD-fmk (pancaspase inhibitor), zIETD-fmk (caspase-8 inhibitor), zAEVD-fmk (caspase-10 inhibitor), zLEHD-fmk (caspase-9 inhibitor), and zDEVD-fmk (caspase-3 inhibitor) were purchased from R&D Systems.

The number of viable cells was determined using a Cell Counting Kit-8 assay according to the manufacturer's instructions (Dojindo). After the incubation of DLD-1 cells for 24 or 48 h with the indicated concentrations of various drugs, kit reagent WST-8 was added to the medium and incubated for a further 2 h. The absorbance of samples (450 nm) was determined using a scanning multiwell spectrophotometer that serves as an ELISA reader.

DNA fragmentation was quantified as the percentage of cells with hypodiploid DNA (sub-G₁). For flow cytometry analysis, cells were exposed to the agents for the indicated times. The cells were then treated with Triton X-100 and RNase A, and their nuclei were stained with propidium iodide before the DNA content was measured using a FACSCalibur flow cytometer and CellQuest software (Becton Dickinson). For all assays, 10,000 cells were counted.

Cells were lysed in a radioimmunoprecipitation assay buffer containing PhosSTOP (Roche Applied Science). Western blot analysis was done as previously described (21), using mouse monoclonal anti–Bcl-2, anti–poly ADP ribose polymerase (Santa Cruz Biotechnology), anti-Bid, anti–caspase-8, anti–caspase-9, anti–caspase-10 (MBL), anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH; HyTest), anti-XIAP (R&D Systems), and anti–β-actin (Sigma). Rabbit monoclonal anti–caspase-3, anti–cleaved caspase-3 (Cell Signaling) and anti-GRP78/Bip (Santa Cruz Biotechnology), rabbit polyclonal anti–acetylated histone H4 (Millipore), anti-histone H4 (Abcam), anti-Bak, anti-Bax, anti-Bcl-X_L, anti–CAAT/enhancer binding protein homologous protein (CHOP/GADD153; Santa Cruz Biotechnology), anti-cIAP-1, anti-survivin (R&D Systems), and anti–death receptor 5 (DR5/TRAIL-R2; Prosci) were supplied as indicated. The signal was detected with an enhanced chemiluminescence Western blot analysis system (GE Healthcare).

For the measurement of ROS production, cells were treated with 20 μmol/L 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, acetylated ester dye (Molecular Probes) for 15 min at 37°C, following a 12-h incubation with agents as indicated. Fluorescence was monitored in the FL-1 channel by a Becton Dickinson FACSCalibur using the CELL Quest software.

Total RNA was isolated from DLD-1 cells treated with agents for 24 h using Sepasol-RNA I (Nacalai Tesque, Inc.) according to the manufacturer's instructions. Total RNA (10 μg) was reversely transcribed to cDNA in a 20 μL reaction volume, with MMTV-reverse transcriptase (Promega), using oligo (dT) primers (Toyobo), according to the manufacturer's instructions. Quantitative real-time reverse transcription-PCR (RT-PCR) was carried out using a RT-PCR system GeneAmp5700 (Applied Biosystems) to quantify the expression level of Bcl-X_L, survivin, cIAP-1, XIAP, CHOP, and DR5 mRNA normalized to 18S rRNA. Real-time RT-PCR primer probes were purchased from Applied Biosystems.

A series of CHOP reporter plasmids and the ER stress response element (ERSE) mutant plasmid pCHOP/mt ERSE (22, 23), and a series of DR5 reporter plasmids and the CHOP mutant plasmid pDR5/mtCHOP (24) were as described previously. DLD-1 cells were seeded at 1 × 10⁵ cells per well in six-well plates and, after 24 h, were transfected with these plasmids or pGVB2 (a vacant control; 1.0 μg) using the DEAE-dextran method and a CellPhect transfection kit (Amersham Pharmacia). After 24 h, the cells were treated with various agents for 24 h and then harvested. Luciferase assays were carried out according to the manufacturer's instructions.

The CHOP and DR5 small interfering RNA (siRNA) duplexes were purchased from Sigma. The DR5 siRNA (sense sequence, 5′-GACCCUUGUGCUCGUUGUCTT-3′; antisense sequence, 5′-GACAACGAGCACAAGGGUCTT-3′) or the CHOP siRNA (sense sequence, 5′-CCAGGAAACGGAAACAGAGTT-3′; antisense sequence, 5′-CUCUGUUUCCGUUUCCUGGTT-3′) was used to knock down the induction of DR5 or CHOP by VPA/15d-PGJ₂, respectively. The LacZ siRNA was used as the siRNA control. One day before the transfection, DLD-1 cells were seeded at 5 × 10⁴ cells per well in six-well plates without antibiotics. The CHOP and/or DR5 siRNAs (25 nmol/L) were transfected into cells using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions. Twenty hours after transfection, the cells were treated with or without VPA/15d-PGJ₂ for 48 h and then harvested.

A full-length human Flag epitope tagged Bcl-X_L cDNA and the pcDNA3 expression vector (neo) were kind gifts from Dr. G. Nuñez (University of Michigan Medical School, Ann Arbor, MI). The pcDNA3-myc-XIAP plasmid constructs and the pcDNA3-myc-tag control (myc) were generous gifts from Dr. R. Takahashi (University of Kyoto, Kyoto, Japan). Transient transfection of DLD-1 cells with overexpression plasmids was conducted using HilyMax transfection reagent (Dojindo) according to the manufacturer's instructions. DLD-1 cells were seeded at 1.5 × 10⁵ cells per well in six-well plates and, after 24 h, were transfected by incubation for 4 h with 4 μg of plasmids and 12 μL of Hilymax for Bcl-X_L and/or XIAP in 2 mL of RPMI 1640 without antibiotics. Twenty-four hours after the transfection, the medium was replaced and the cells were treated with or without VPA/15d-PGJ₂ for 48 h and then harvested.

Male BALB/c nu/nu mice were purchased from Shimizu Co., Ltd. On day 0, 7.5 × 10⁶ DLD-1 cells were resuspended in 1× phosphate-buffered NaCl solution, mixed with 1× Matrigel (BD Biosciences), and inoculated s.c. into the upper-right flank of each mouse. On day 3, when the tumors reached 6 to 8 mm in diameter, the mice were randomly divided into four groups (five mice per group) and treatment was initiated. Mice were injected everyday (i.p.) with diluent (10% DMSO in PBS, control group), VPA (100 mg/kg/d), 15d-PGJ₂ (5 mg/kg/d) or a combination for 25 d. Tumors were measured with a sliding caliper twice weekly and the volumes (mm³) were calculated according to the formula: V = A × B² × 0.52, in which A is the largest superficial diameter and B is the smallest superficial diameter. On day 28, the tumors were excised from the euthanized mice. All experiments and procedures were done in accordance with the Institutional Care Use Committee guidelines.

Statistical evaluation of the data was done using the Student's t test for simple comparison between groups and treatments. P < 0.05 was considered statistically significant. Characterization of synergistic interactions was done using median dose effect analysis in conjunction with a commercially available software program (CalcuSyn, Biosoft).

To investigate the antiproliferative effects of 15d-PGJ₂ or HDIs alone, we assessed the viability of DLD-1 cells treated with the indicated concentrations of agents for 24 or 48 hours. We observed that each agent showed minimally antiproliferative effects on DLD-1 cells in a dose- and time-dependent manner (Fig. 1A). Interestingly, cotreatments with 15d-PGJ₂ and HDIs markedly inhibited the cell viability compared with that of cells treated with either 15d-PGJ₂ or HDIs alone (Fig. 1B). Moreover, the combination index values for both 15d-PGJ₂ and SAHA or VPA were <1.0, indicating synergistic cell death–inducing efficacy (Fig. 1B).

To clarify the mechanisms of synergistic inhibitory effects on cell viability by incubation with the combinations of 15d-PGJ₂ and HDIs, we examined the effects of the combined treatments on apoptosis by measuring the sub-G₁ population. Cotreatments with 15d-PGJ₂ and HDIs drastically induced apoptosis in a dose-dependent manner (Fig. 1C). As shown in Fig. 1D, the pan-caspase inhibitor zVAD-fmk, caspase-3, caspase-9, caspase-10, and caspase-8 inhibitors efficiently blocked apoptosis induced by cotreatments with 15d-PGJ₂ and HDIs, indicating that the induced apoptosis is mediated through the activation of both the intrinsic pathway (mitochondrial pathway) and the extrinsic pathway (death receptor pathway). It was reported that 15d-PGJ₂– or HDI-induced lethality is associated with ROS production (18, 25). Pretreatment with the antioxidant NAC, which functions as a precursor of intracellular glutathione, significantly inhibited apoptosis induced by the combined treatments with 15d-PGJ₂ and HDIs (Fig. 1D). On the other hand, a ROS inducer buthionine sulfoximine, which causes the depletion of intracellular glutathione, attenuated the apoptosis-inhibitory effect of NAC. These results suggest that cotreatments with 15d-PGJ₂ and HDIs synergistically induce caspase- and ROS-dependent apoptosis.

To elucidate the mechanism of apoptosis induced by VPA/15d-PGJ₂ cotreatment, we carried out Western blot analysis. VPA at 0.5 mmol/L for 6 to 36 hours clearly induced acetylated histone H4 and the treatment with VPA for 12 hours increased acetylated histone H4 in a dose-dependent manner (Fig. 2A). Interestingly, VPA/15d-PGJ₂ cotreatment for 12 hours significantly induced acetylated histone H4 compared with VPA alone. VPA and 15d-PGJ₂ cooperated in the activation of caspases and the cleavage of Bid and PARP (Fig. 2B). On the other hand, VPA/15d-PGJ₂ cotreatment significantly decreased the expression of antiapoptotic molecules such as a prosurvival Bcl-2 family member Bcl-X_L and inhibitor of apoptosis proteins family members survivin, cIAP-1, and XIAP in dose- and time-dependent manners compared with 15d-PGJ₂ alone (Fig. 2C and D).

ER stress was reported to be an event downstream of intense or persistent oxidative stress (26) and to induce ER stress markers such as a glucose-regulated protein GRP78/Bip and a proapoptotic molecule CHOP (27). We previously showed that an apoptosis-inducing membrane receptor, DR5, is a downstream target of CHOP (24). In the present study, VPA/15d-PGJ₂ cotreatment significantly increased GRP78/Bip, CHOP, and DR5 expression in dose- and time-dependent manners compared with 15d-PGJ₂ alone (Fig. 2C and D). Taken together, these results suggest that VPA/15d-PGJ₂ cotreatment causes apoptosis through the downregulation of antiapoptotic molecules and the upregulation of CHOP and DR5, and that the activation of both the intrinsic and extrinsic pathways are involved in the induced apoptosis.

Next, we further investigated whether ROS generation could relate to the induced apoptosis. VPA/15d-PGJ₂ cotreatment markedly increased the ROS level, which was blocked by NAC pretreatment but not by zVAD-fmk (Fig. 3A). Moreover, neither NAC nor zVAD-fmk pretreatment decreased acetylated histone H4 induced by VPA/15d-PGJ₂ cotreatment, suggesting that HDAC inhibition is an upstream event of oxidative injury– and subsequent caspase-dependent apoptosis caused by VPA/15d-PGJ₂ (Fig. 3B). NAC and zVAD-fmk inhibited VPA/15d-PGJ₂–induced activation of caspases and cleavage of PARP (Fig. 3C). VPA/15d-PGJ₂ cotreatment downregulated the expression of Bcl-X_L and XIAP and upregulated that of CHOP and DR5 at both the protein and mRNA levels, and these alterations were restored by NAC pretreatment but not by zVAD-fmk (Fig. 3C and D). Unlike these genes, decreased cIAP-1 expression was recovered not only by NAC pretreatment but also by zVAD-fmk, raising the possibility that the downregulation of cIAP-1 is a consequence of caspase activation and/or apoptosis induced by VPA/15d-PGJ₂ cotreatment. Decreased expression of survivin was not significantly restored by NAC or zVAD-fmk (Fig. 3C and D). These results indicate that Bcl-X_L, XIAP, CHOP, and DR5 play important roles in the oxidative injury–mediated apoptosis induced by VPA/15d-PGJ₂ cotreatment.

SAHA at 1.5 μmol/L for 6 to 24 hours clearly induced acetylated histone H4 and the treatment with SAHA for 12 hours increased acetylated histone H4 in a dose-dependent manner (Supplementary Fig. S1A). Additionally, SAHA/15d-PGJ₂ cotreatment significantly induced ROS generation; decreased the expression of Bcl-X_L and XIAP; and increased that of GRP78/Bip, CHOP, and DR5, which were inhibited by NAC pretreatment (Supplementary Fig. S1B and D). Neither NAC nor zVAD pretreatment decreased acetylated histone H4 induced by SAHA/15d-PGJ₂ cotreatment (Supplementary Fig. S1C). These results suggest that SAHA has pretty similar effects to those by VPA when combined with 15d-PGJ₂, although VPA has more clearly shown the synergistic effects with 15d-PGJ₂ than SAHA. Taken together, these data indicate that significantHDAC inhibition is the first step in the oxidative injury– and caspase-dependent apoptosis caused by HDIs/15d-PGJ₂.

To clarify whether the downregulation of the Bcl-X_L and XIAP promoter was responsible for the inhibition of these mRNAs, we examined the promoter activities of Bcl-X_L and XIAP. The promoter activities of these genes were not decreased by VPA/15d-PGJ₂ cotreatment (data not shown). Therefore, this led us to investigate whether a posttranscriptional mechanism could be involved in the regulation of Bcl-X_L and XIAP by analyzing the decay rate of these mRNAs. VPA/15d-PGJ₂ cotreatment decreased the decay rate of XIAP mRNA but not that of Bcl-X_L, indicating that these mRNAs are differentially regulated (Fig. 4A). Transient transfection of either Bcl-X_L or XIAP showed that both Bcl-X_L and XIAP partially protected against VPA/15d-PGJ₂–mediated apoptosis (Fig. 4B). Moreover, the cells cotransfected with Bcl-X_L and XIAP were found to be significantly more resistant to VPA/15d-PGJ₂–induced apoptosis compared with those transfected with either Bcl-X_L or XIAP. These results suggest that VPA/15d-PGJ₂ cotreatment, at least partially, causes apoptosis through the downregulation of Bcl-X_L and XIAP.

We next carried out a luciferase assay using reporter plasmids containing the CHOP or DR5 promoter. VPA/15d-PGJ₂ cotreatment synergistically stimulated the promoter activity from pCHOP/-150 or pDR5/-347, which was attenuated by NAC pretreatment (Fig. 4C). Moreover, the mutation of ERSE abolished the activation of the CHOP promoter by VPA/15d-PGJ₂. The promoter region from pDR5/-252 or pDR5/-347 with the CHOP mutation did not respond to VPA/15d-PGJ₂. These results suggest that synergistic ROS generation induced by VPA/15d-PGJ₂ cotreatment is an upstream event of ER stress, leading to the increase of the promoter levels of CHOP and DR5. Additionally, the siRNA of DR5 or CHOP almost completely decreased DR5 or CHOP protein and significantly attenuated apoptosis in DLD-1 cells treated with VPA/15d-PGJ₂, respectively, compared with control LacZ siRNA (Fig. 4D). Importantly, concomitant with CHOP reduction by CHOP siRNA, VPA/15d-PGJ₂–induced DR5 expression was at least partially suppressed in CHOP siRNA–transfected cells when compared with control LacZ siRNA–transfected cells. No significant difference in the attenuation of apoptosis among CHOP-, DR5- and CHOP/DR5 siRNA–transfected cells treated with 15d-PGJ₂/VPA was detected. Moreover, no restoration of decreased Bcl-X_L and XIAP by CHOP or DR5 siRNA was observed. Taken together, these results indicate that CHOP is involved in 15d-PGJ₂/VPA–induced apoptosis, mainly due to DR5 induction, and further support the notion that CHOP-mediated DR5 induction is an important step in the extrinsic pathway activated by oxidative injury and subsequent ER stress.

To determine whether the effects of VPA/15d-PGJ₂ cotreatment on the induction of apoptosis or ROS generation could be observed more generally, other malignant tumor cell lines such as HaCaT, K562, and HT29 were similarly assayed. VPA/15d-PGJ₂ caused 74%, 44%, and 43% apoptosis in HaCaT, K562, and HT29 cells, respectively, which were effectively attenuated by NAC as seen in DLD-1 cells (Fig. 5A). VPA/15d-PGJ₂ cotreatment significantly downregulated the expression of Bcl-2, Bcl-X_L, cIAP-1 and XIAP in HaCaT cells, cIAP-1 and XIAP in K562 cells and survivin, cIAP-1 and XIAP in HT29 cells, and upregulated the expression of CHOP and DR5 in HaCaT cells, compared with VPA or 15d-PGJ₂ alone. These alterations were inhibited by NAC. Additionally, the combined treatment led to a significant increase in the level of ROS in these cells, which was inhibited by NAC (Fig. 5B). Moreover, we found that this combination did not lead to significant apoptosis in normal peripheral blood mononuclear cells (data not shown). These results suggest that VPA/15d-PGJ₂ cotreatment causes apoptosis in various malignant tumor cells with low toxicity to normal cells, correlating with the alteration of apoptosis-related molecules through ROS generation.

Based on the in vitro antitumor efficacy of the combined treatment with VPA and 15d-PGJ₂, we evaluated the antitumor activity of the combined therapy in a nude mice xenograft model inoculated with DLD-1 cells. Although either VPA or 15d-PGJ₂ treatment alone did not inhibit the growth of DLD-1 tumors, the combination therapy significantly suppressed the growth compared with the control on day 18 and later (Fig. 6A and B). These results suggest that a combination of VPA and 15d-PGJ₂ is more effective than a VPA or 15d-PGJ₂ monotherapy in vivo as observed in vitro.

Recent preclinical studies have suggested that despite the promising anticancer effects of HDIs alone, they also enhance the antitumor activities of a large number of anticancer agents, including conventional chemotherapeutic and molecular-targeted drugs (28). It is likely that HDIs have a broader capacity to lower the threshold for tumor cells to undergo apoptotic cell death triggered by other agents. Therefore, to enhance the antitumor activity of HDIs and/or to overcome resistance to HDIs, mechanism-based therapeutic strategies using a combination of HDIs and other agents should be developed.

It has been shown that ROS-oxidation-reduction pathways are important mechanisms of HDI-induced transformed cell death as a monotherapy or combination therapies with other agents. Xu et al. (29) suggested that the downregulation of the endogenous ROS scavenger thioredoxin is associated with SAHA-induced ROS-dependent apoptosis. Moreover, ROS generation plays a crucial role in synergistic apoptosis caused by the combinations of HDIs and various compounds such as the proteasome inhibitor bortezomib, the purine analogue fludarabine, the MAP/ERK kinase 1/2 inhibitor PD184352, the alkyl-lysophospholipid perifosine, and the multiple receptor tyrosine kinase inhibitor AEE788 (30–34). The present study suggests that ROS generation through HDAC inhibition leads to caspase-dependent apoptosis caused by HDIs/15d-PGJ₂.

Fajas et al. (35) reported that VPA has a synergistic effect with the synthetic PPARγ ligands thiazolidinediones in the activation of PPARγ target genes. Therefore, VPA may render PPARγ permissive to activation by 15d-PGJ₂. However, pretreatment with the PPARγ irreversible antagonist GW9662 did not inhibit apoptosis induced by coadministration of 15d-PGJ₂ and HDIs, and unlike 15d-PGJ₂, the thiazolidinediones pioglitazone and ciglitazone failed to increase HDI-induced apoptosis even at a high concentration of 50 μmol/L, suggesting that the apoptosis induced by VPA/15d-PGJ₂ is independent of PPARγ activation (Supplementary Fig. S2A and B).

We showed that VPA/15d-PGJ₂ cotreatment drastically decreased Bcl-X_L and XIAP expression at the mRNA level through ROS generation. VPA/15d-PGJ₂ caused the destabilization of XIAP but not Bcl-X_L mRNA, indicating that these mRNAs are differentially regulated. Adenylate/uridylate-rich elements, often containing one to several copies of the consensus sequence AUUUA, are critical cis-acting regulatory motifs located in the 3′-untranslated region of many mRNAs and are targets for trans-acting proteins to modulate mRNA stability (21). We found that human XIAP mRNA contains 20 separate AUUUA repeats in the 3′-untranslated region, whereas Zhang et al. (36) reported that there are three AUUUA elements on the 3′-untranslated region of Bcl-X_L mRNA. Therefore, this raises the possibility that VPA/15d-PGJ₂ may affect XIAP mRNA stability through adenylate/uridylate-rich element–binding proteins that are different to those of Bcl-X_L.

We previously reported that treatment with 18 μmol/L 15d-PGJ₂ for 24 hours induces DR5 expression through mRNA stabilization in prostate cancer PC3 cells (37). Su et al. (38) showed that treatment with 30 μmol/L 15d-PGJ₂ for 24 hours triggers CHOP gene expression, resulting in the activation of the DR5 promoter in colon cancer HCT116 cells. These differences may depend on the concentrations of 15d-PGJ₂ used and/or particular properties of different cell lines. In the present study, VPA significantly enhanced DR5 promoter activity through the CHOP binding site and enhanced CHOP promoter activity through the ERSE in DLD-1 cells treated with 8 μmol/L 15d-PGJ₂, which were inhibited by NAC pretreatment. Taken together with the result that VPA/15d-PGJ₂ cotreatment for 12 hours cooperatively upregulated acetylated histone H4, these data suggest that ROS generation caused by HDAC inhibition as an upstream event of ER stress finally leads to an increase in the promoter levels of CHOP and DR5.

Based on the early preclinical data, VPA and SAHA were selected as candidates of HDIs for clinical trial. Kelly et al. (39) reported that in patients with advanced hematologic malignancies treated with oral-administered SAHA at 400 mg/d, the average maximum plasma concentrations of SAHA were 2.0 to 2.5 μmol/L. Münster et al. (12) suggested that the maximum tolerated dose and recommended phase II dose of VPA was 140 mg/kg/d in a phase I trial of VPA followed by the topoisomerase II inhibitor epirubicin in advanced solid tumors. Moreover, they showed that when patients with advanced solid tumors received a daily dose of 15 or 30 mg/kg, VPA reached serum concentrations between 50 and 130 μg/mL (0.3-0.8 mmol/L), which is the therapeutic range for antiepileptic treatment. Our in vitro data indicated that VPA at 0.5 mmol/L and SAHA at 1.5 μmol/L, which are clinically achievable concentrations, drastically induced apoptosis in DLD-1 cells when combined with 15d-PGJ₂.

Regarding antitumor activities of 15d-PGJ₂ in vivo, it has been reported that 15d-PGJ₂ given at 2.5 mg/kg i.p. synergistically improves the antiangiogenic and antitumor effects of the thrombospondin peptide ABT510 in a bladder cancer xenograft (40) and that 15d-PGJ₂ given at 1 mg/kg i.p. enhances docetaxel antitumor activity against a non–small cell lung cancer xenograft (41). Recently, Shin et al. (42) showed that 15d-PGJ₂ given at a dose of 5 mg/kg through the tail vein every 3 days significantly decreased tumor growth in a mouse CT-26 s.c. colorectal cancer model or HL-60 human leukemia xenograft model. They suggested that mice treated with 15d-PGJ₂ did not show signs of wasting and that body weights after 14 days were not significantly different from controls. Moreover, they assessed the toxicity of 15d-PGJ₂ in mice by checking the level of glutamate-oxaloacetate transaminase and glutamate-pyruvate transaminase. They concluded that the dosage used in their experiment was safe based on the results for blood chemistry. In the present study, we showed that combination therapy with VPA at a dose of 100 mg/kg and 15d-PGJ₂ at a dose of 5 mg/kg significantly suppressed the growth of a DLD-1 xenograft compared with the control, whereas either VPA or 15d-PGJ₂ treatment alone did not inhibit the tumor growth. We observed no signs of toxicities such as skin rush, diarrhea, and wasting in mice treated with VPA at 100 mg/kg/d, 15d-PGJ₂ at 5 mg/kg/d, or VPA and 15d-PGJ₂ in combination, compared with control mice. Based on our in vivo study, clinical application of 15d-PGJ₂ can be expected. However, a synthetic method to produce sufficient amount of 15d-PGJ₂ for preclinical or clinical studies has not been established (43).

In conclusion, we have shown that the cotreatments with HDIs and 15d-PGJ₂ synergistically induce apoptosis through HDAC inhibition and subsequent ROS generation and ER stress. The cotreatment with VPA and 15d-PGJ₂ drastically induces apoptosis in various malignant tumor cells in vitro and inhibits tumor growth in vivo. These observations raise the possibility that the combined treatment with clinically relevant HDIs and the natural endogenous PPARγ ligand 15d-PGJ₂ could prove to be promising for the treatment of a broad spectrum of malignant tumors in the clinic.

No potential conflicts of interest were disclosed.

We thank R. Nakanishi and H. Watanabe for their experimental help.

Grant Support: Grant-in-Aid from the Japanese Ministry of Education, Culture, Sports, Science and Technology.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.