The HHUA human endometrial cancer cell line was obtained from Riken (Ibaraki, Japan). The Ishikawa human endometrial cancer cell line was kindly provided by Dr Masato Nishida (Tsukuba University, Ibaraki, Japan). The HEC-59 human endometrial cancer cell line was obtained from the American Type Culture Collection (Manassas, VA, USA). The cells were maintained as monolayers at 37°C in 5% CO₂/air in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Rockville, MD, USA) containing 10% heat-inactivated fetal bovine serum (FBS; Omega, Tarzana, CA, USA).

15d-PGJ₂ was obtained from Enzo Life Sciences (Plymouth Meeting, Montgomery County, PA, USA), and prepared as a 20 mg/ml stock solution in dimethyl sulfoxide (DMSO). The stock solution was stored in aliquots at −20°C.

The cell proliferation and cell viability were determined in 96-well plates by a modified methylthiazol tetrazolium (MTT) assay using WST-1 (Roche Diagnostics, Penzberg, Germany) following the manufacturer’s instructions. We distributed 5×10³ cells in DMEM supplemented with 10% FBS into each well of a 96-well flat-bottomed microplate (Corning, Inc., New York, NY, USA) and incubated them overnight. The medium was then removed, and the cells were incubated for 48 h with 100 μl of experimental medium containing various concentrations of 15d-PGJ₂. Thereafter, 10 μl of WST-1 dye was added to each well, and the cells were further incubated for 4 h. All experiments were performed in the presence of 10% FBS. Cell proliferation was evaluated by measuring the absorbance at 540 nm. Data were calculated as the ratio of the values obtained for the 15d-PGJ₂-treated cells to those for the untreated controls.

The cell cycle was analyzed by flow cytometry after 2 days of culturing. Cells (5×10⁴) were exposed to 15d-PGJ₂ in 6-well flat-bottomed plates for 48 h. Analysis was performed immediately after staining using the CellFIT program (Becton-Dickinson, San Jose, CA, USA), whereby the S phase was calculated using an RFit model.

Cells were plated and grown overnight until they reached 80% confluence and then treated with 15d-PGJ₂. After 48 h, detached cells in the medium were collected, and the remaining adherent cells were harvested by trypsinization. The cells (1×10⁵) were washed with PBS and resuspended in 250 μl of binding buffer (Annexin V-FITC kit; Becton-Dickinson) containing 10 μl of 20 μg/ml PI and 5 μl of Annexin V-FITC, which binds to phosphatidylserine translocated to the exterior of the cell membrane early in the apoptotic pathway as well as during necrosis. After incubation for 10 min at room temperature in a light-protected area, the samples were analyzed on a FACSCalibur flow cytometer (Becton-Dickinson). FITC and PI emissions were detected in the FL-1 and FL-2 channels, respectively. For each sample, data from 30,000 cells were recorded in list mode on logarithmic scales. Subsequent analysis was performed with CellQuest software (Becton-Dickinson).

Cells were prepared for FACS analysis as described above and stained using a Mitocapture Apoptosis Detection kit obtained from BioVision (Palo Alto, CA, USA) with a fluorescent lipophilic cationic reagent that assesses mitochondrial membrane permeability, according to the manufacturer’s recommendations.

Total RNA was extracted from the 15d-PGJ₂-treated and untreated HHUA cells using an RNeasy mini kit (Qiagen, Valencia, CA, USA) in accordance with the manufacturer’s instructions. Prior to hybridization, the quantity and quality of the total RNA were evaluated using a spectrophotometer and a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA), respectively. Cy3-labeled cRNA targets were generated using a Low RNA Input Fluorescent Linear Amplification kit (Agilent Technologies). A human 44 K oligoarray was used for hybridization, in accordance with the manufacturer’s recommendations (Agilent Technologies). A laser confocal scanner (Agilent Technologies) was used to measure signal intensities in the expression microarray analysis. Feature Extraction software (Version 9.1; Agilent Technologies) with the manufacturer’s recommended settings was applied for the microarray image analysis. Analysis of the microarray images was performed with GeneSpring 7.3.1 software (Agilent Technologies). For comparison among multiple arrays, probe set data were median-normalized/chip. The data were then centered across the genes in 6 normal controls, followed by filtering based on a signal intensity of ≥100, and contained no flagged values. Among these differentially expressed genes, those designated as ‘upregulated’ were overexpressed >2-fold in comparison with the controls (P<0.05), whereas those designated as ‘downregulated’ were underexpressed <0.75-fold compared with the controls (P<0.05). Annotations including chromosomal loci were provided by Agilent Technologies.

For Gene Ontology (GO) analysis, differentially expressed genes were defined as those with a >2-fold increase or decrease in expression relative to the controls. GO term enrichment in the upregulated or downregulated gene sets was assessed using the GOstat web tool (13).

Cells were washed twice in PBS, suspended in lysis buffer [50 mM Tris (pH 8.0), 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1% NP-40, phenylmethylsulfonyl fluoride at 100 μg/ml, aprotinin at 2 μg/ml, pepstatin at 1 μg/ml and leupeptin at 10 μg/ml], and placed on ice for 30 min. After centrifugation at 15,000 × g for 15 min at 4°C, the suspension was collected. Protein concentrations were quantified using the Bio-Rad protein Assay Dye Reagent Concentrate (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer’s recommendations. Whole-cell lysates (40 μg) were resolved by SDS-polyacrylamide gel electrophoresis on a 4–15% gel, transferred onto a polyvinylidene difluoride membrane (Immobilon; Amersham, Arlington Heights, IL, USA), and probed sequentially with antibodies against anterior gradient homolog 3 (AGR3; 1:1,000; GeneTex, Irvine, CA, USA), aldo-keto reductase family 1 member C1 (AKR1C1; 1:1,000; GeneTex), aldo-keto reductase family 1 member C3 (AKR1C3; 1:1,000; ProteinTech, Chicago, IL, USA), α-1-microglobulin/bikunin precursor (AMBP; 1:1,000; Abnova, Taipei, Taiwan), complement component 3a receptor 1 (C3AR1; 1:1,000; Abnova), chondroadherin (CHAD; 1:1,000; Avia Systems Biology, San Diego, CA, USA), Fer3-like (Drosophila) (FERDL3; 1:1,000; Avia Systems Biology), ferritin, light polypeptide (FTL; 1:1,000; GeneTex), galactose-3-O-sulfotransferase 3 (GAL3ST3; 1:1,000; Avia Systems Biology), glutamate-cysteine ligase, modifier subunit (GCLM; 1:1,000; Abnova), heme oxygenase (decycling) 1 (HMOX1; 1:1,000; Abnova), intercellular adhesion molecule 4 (ICAM4; 1:1,000; Abnova), potassium voltage-gated channel, shaker-related subfamily, β member 1 (KCNAB1; 1:1,000; Osenses, Keswick, Australia), mitochondrial ribosomal protein L37 (MRPL37; 1:1,000; ProteinTech), nitric oxide synthase 2A (NOS2A; 1:1,000; Applied Biological Materials, Kampenhout, Belgium), phosphorylated eukaryotic translation initiation factor 4E (p-eIF4E; 1:1,000; Bioworld Technology, Minneapolis, MN, USA), pirin (PIR; 1:1,000; Avia Systems Biology), tripartite motif-containing 16 (TRIM16; 1:1,000; Avia Systems Biology), thioredoxin reductase 1 (TXNRD1; 1:1,000; ProteinTech), UDP glucuronosyltransferase 1 family, polypeptide A6 (UGT1A6; 1:1,000; LifeSpan Biosciences, Seattle, WA, USA) and GAPDH monoclonal antibody (mAb) (1:10,000; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). The blots were developed using an enhanced chemiluminescent (ECL) kit (Amersham). Band intensity was measured using the public domain Image program ImageJ version 1.44, and fold increase in expression as compared with control, untreated cells was calculated.

Data are presented as the means ± SD of representative experiments and were analyzed by the Bonferroni-Dunn test using StatView 4.5 software (Abacus Concepts, Berkeley, CA, USA). A P-value <0.05 was considered to indicate a statistically significant difference.

The antitumor effects of 15d-PGJ₂ on 3 endometrial cancer cell lines in vitro were examined using a WST-1 assay of the 2-day exposure to 15d-PGJ₂. Significant inhibitory effects of 15d-PGJ₂ on the cell growth were observed in all 3 endometrial cancer cell lines (Ishikawa, HHUA and HEC-59) (Fig. 1).

We then investigated whether 15d-PGJ₂ would lead to the induction of apoptosis and/or cell cycle arrest in the endometrial cancer cells (Table I). 15d-PGJ₂ led to an increase in the sub G0/G1 apoptotic cell population and the cell population in the G2/M phase of the cell cycle compared to treatment with the vehicle alone, with a concomitant decrease in the proportion of cells in the S phase.

To assess the ability of the endometrial cancer cells to undergo apoptosis in response to 15d-PGJ₂ exposure and to distinguish between the different types of cell death, we double-stained the 15d-PGJ₂-treated cells with Annexin V and PI and analyzed the results using flow cytometry. Annexin V binding combined with PI labeling was performed for the distinction of early apoptotic (Annexin V⁺/PI⁻) and necrotic (Annexin V⁺/PI⁺) cells. At increasing doses of 15d-PGJ₂, a simultaneous increase in both the Annexin V⁺/PI⁻ fraction (early apoptotic) and Annexin V⁺/PI⁺ (regarded as necrotic) subpopulations was detected (Table II).

It has been shown that the loss of MTP occurs prior to nuclear condensation and caspase activation and is linked to cytochrome c release in many, but not all, apoptotic cells (14,15). It was found that the treatment of endometrial cancer cells with 15d-PGJ₂ resulted in the loss of MTP (Table II).

In order to identify potential and novel target genes responsive to the anticancer effects in 15d-PGJ₂-treated endometrial cancer cells, we examined the global changes in gene expression in the HHUA cells following treatment with 10 μM of 15d-PGJ₂ for 48 h (Tables III and IV). Of the 44,000 genes, GO analysis was carried out on the genes upregulated and downregulated by the treatment (Tables V and VI).

To elucidate the common mechanism of action of 15d-PGJ₂ in endometrial cancer, we examined the effects of 15d-PGJ₂ on the expression of 20 proteins that were selected from the cDNA microarray data in 3 endometrial cancer cell lines using western blot analysis (Fig. 2 and Table VII). 15d-PGJ₂ markedly upregulated the levels of AKR1C3 and downregulated the levels of AGR3 and NOS2A proteins in all 3 endometrial cancer cell lines.