Tissue specimens were obtained from 34 patients without pre-surgical chemotherapy or radiotherapy at the Department of Gastroenterological Surgery, Hangzhou First Hospital, School of Clinical Medicine, Nanjing Medical University between January 2009 and December 2013. The study was approved by our hospital review board, and each patient provided informed consent. Fresh tissue specimens and paraffin blocks were used to assess gene expression.

Total cellular RNA was isolated from tissue samples or cells using an RNeasy Mini kit (Biomed, Beijing, China) and reversely transcribed into first-strand cDNA using a Takara reverse transcription kit (Takara, Dalian, China) and oligo (dT)15 primers (Takara) according to the manufacturer’s instructions. The resultant cDNA was then used for qPCR amplification of PDCD2 and p53 expression, and GAPDH mRNA was used as a control. PDCD2 primers were 5′-CTGTGGAGCTGGGCTTCGCC-3′ and 5′-CAGCAGGAAGGAGAGCGGGC-3′. p53 primers were 5′-ACTCCAGCCACCTGTAGTCCAAAAAGGGTC-3′ and 5′-GACCCTTTTTGGACTACAGGTGGCTGGAGT-3′. GAPDH primers were 5′-AGAAGGCTGGGGCTCATTTG-3′ and 5′-AGGGGCCATCCACAGTCTTC-3′. Amplification of PDCD2, p53 and GADPH mRNA was performed with one cycle at 95°C for 10 min and 40 cycles of 95°C for 15 sec and 60°C for 60 sec. Calculation of the relative expression of each transcript was performed using the 2^−ΔΔCt method.

Tissues and cells were lysed in a lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA and 1% Triton-X100) containing a protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO, USA). Cellular protein was then quantified using the BCA protein assay kit (Beyotime, Beijing, China). Equivalent amounts of protein samples (20 μg) were separated using 12% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore Corp., Billerica, MA, USA). Anti-PDCD2 (sc-377250) and anti-β-actin (sc-130301) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA), and cell cycle checkpoint-regulated proteins, including anti-ataxia telangiectasia mutated (ATM), anti-p53, anti-checkpoint kinase (Chk)1 and anti-Chk2 were purchased from Santa Cruz Biotechnology. Anti-phospho-S345-Chk1 and anti-phospho-T68-Chk2 antibodies were from Cell Signaling Technology (Danvers, MA, USA). Each specific antibody binding was detected with horseradish peroxidase (HRP)-conjugated respective secondary antibodies (Amersham Biosciences, Amersham, UK) and enhanced chemiluminescence (ECL) solutions (Amersham Biosciences).

Tumor tissues were fixed in 4% para-formaldehyde for 24 h, embedded in paraffin and then sectioned into 4-μm sections for immunohistochemistry. Briefly, endogenous peroxidase activity was blocked by incubating sections in 3% hydrogen peroxide for 30 min. Antigen retrieval was performed in citrate buffer (10 mM, pH 6.0) for 30 min at 95°C in a pressure cooker. After that, the sections were blocked in 1.5% blocking serum in phosphate-buffered saline (PBS) for 2 h at room temperature and incubated with an anti-PDCD2 antibody (1:200 dilution) or anti-p53 antibody (1:200 dilution; both from Santa Cruz Biotechnology) overnight at 4°C in a moist chamber. The next day sections were washed with PBS and incubated with a biotinylated secondary antibody at 37°C for 2 h before exposure to a streptavidin complex (HRP; Beyotime). Positive reaction was visualized using 3,3′-diami-nobenzidine tetrahydrochloride (DAB; Beyotime) followed by counterstaining with hematoxylin (Beyotime). The stained sections were reviewed and scored independently under a light microscope by two researchers. Positive staining was defined as 25% or more of tumor cells staining positively.

Gastric cancer MKN28 cells (p53 mutant) and MKN45 cells (p53 wild-type) were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA) and grown in RPMI-1640 medium (HyClone, Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA, USA) and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin) in a humidified incubator with 5% CO₂ at 37°C.

Plasmids carrying PDCD2 or p53 cDNA, namely pCDNA-3.1-PDCD2 or pCDNA-3.1-p53, respectively, were obtained from Dr Cong Jie (China Medical University). p53 shRNA (cat # sc-29435) plasmid was obtained from Santa Cruz Biotechnology. For gene transfection, gastric cancer MKN28 and MKN45 cells were grown overnight and transfected with pCDNA-3.1-PDCD2 using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Plasmid pCDNA-3.1-p53 was transfected into MKN28-PDCD2 cells and p53 shRNA was transfected into MKN45-PDCD2 cells using Lipofectamine 2000.

[Methyl-³H]thymidine was obtained from Sigma-Aldrich. Cells (1,000 cells/well) following gene transfection were seeded at least in quadruplicate in 96-well plates, grown for 4 days and then incubated with 20 mCi/well [methyl-³H] thymidine for 16 h. Subsequently, the cells were washed with ice-cold PBS and lysed with ice-cold 10% trichloroacetic acid (TCA) in PBS for 20 min at 4°C. Ice-cold 0.2 M NaOH was then added to the cells and incubated for 10 min at −20°C. The cell lysates were then collected, universal scintillation cocktail (2.5 ml) was added and the incorporation of [³H]thymidine into DNA was assessed using a liquid scintillation counter (Xi’an Nuclear Instrument Factory, Xi’an, China) to determine the disintegrations per minute (dpm) values.

To detect cell cycle distribution, the cells following 48 h of gene transfection were harvested, suspended in PBS and fixed with 70% ethanol at 4°C. After centrifugation (1,500 × g for 5 min), the supernatants were discarded, cellular DNA was stained with 10 μM propidium iodide (PI; Keygen, Nanjing, China) and samples were analyzed by a FACSCalibur flow cytometer (BD Biosciences, Baltimore, MD, USA).

Apoptosis was determined using an apoptosis detection kit (Keygen). Briefly, cells following 48 h of gene transfection were collected, washed twice in ice-cold PBS and then resuspended in binding buffer at a density of 1×10⁶ cells/ml. The cells were simultaneously incubated with fluorescein-labeled Annexin V and PI for 20 min. The mixture was then analyzed using a FACSCalibur (BD Biosciences).

Actively replicating cells at the beginning of each hour of the S phase were first labeled with the thymidine analog, 5-iodo-2′-deoxyuridine (IdU; 50 μM, Sigma-Aldrich) for 40 min, washed three times with PBS and then labeled with 5-chloro-2′-deoxyuridine (CldU; 100 μM, Sigma-Aldrich) for 40 min. The IdU and CldU incorporated into replicating DNA were then detected with red or green fluorescent antibodies, respectively. Specifically, slides were treated with 70% ethanol, washed in PBS, denatured in 2.5 M HCl for 30 min, permeabilized in 0.25% Triton X-100 for 5 min and blocked with 1% bovine serum albumin. The slides were incubated at room temperature with the following antibodies: i) a mouse anti-bromodeoxyuridine antibody at a dilution of 1:500 (detects IdU; Sigma-Aldrich); ii) an AlexaFluor 488-conjugated anti-mouse antibody at a dilution of 1:1,000 (Invitrogen); iii) a rat anti-bromodeoxyuridine antibody at a dilution of 1:2,000 (detects CldU; Santa Cruz Biotechnology) and iv) an AlexaFluor 633-conjugated anti-rat antibody at a dilution of 1:1,000 (Invitrogen). After counter-staining with DAPI (1 μg/ml, KeyGen), images were captured under an Olympus CX71 fluorescence microscope (Olympus, Tokyo, Japan).

The animal study was approved by the ethics committee of our University. According to a previous study (11), male haired p53^−/− mice with a C57BL/6J genetic background from the Jackson Laboratory (Bar Harbor, ME, USA) were mated with female hairless SKH-1 mice (p53 wild-type) to obtain male and female hairless congenic p53-deficient mice. The hairless p53^−/− mice were intercrossed to obtain homozygous p53-deficient mice and their wild-type littermates. Then, UV lamps (FS72T12-UVB-HO, National Biological Corp., Twinsburg, OH, USA) were used to expose 30 p53^−/− and 30 p53^+/+ SKH-1 mice to 180 mJ/cm² UVB (280–320 nm; 75–80% of total energy) twice/week for 20 weeks. This irradiation induced skin tumors in 100% of irradiated mice. At the end of the experiment, tumor-bearing p53^−/− or p53^+/+ SKH-1 mice were randomly divided into two groups of 15 each for PDCD2 modulation. Specifically, pCDNA3.1-PDCD2 carrying PDCD2 cDNA was injected into the tail vein of each mouse. Two injections were administered at 8 a.m. and 8 p.m. for 3 days. Tumor number was recorded once every 2 weeks and plotted against time (in weeks). At the end of 25 weeks, the experiment was terminated. All of the animals were sacrificed, and skin and tumor samples were harvested.

To assess expression of p53 and PDCD2 proteins in mouse tissues, the double-immunofluorescence technique with specific antibodies against p53 (rabbit IgG, sc-6243, Santa Cruz Biotechnology) and PDCD2 (mouse IgG1, sc-377250, Santa Cruz Biotechnology) was performed. Anti-rabbit AlexaFluor^® 488 IgG or anti-mouse AlexaFluor^® 594 IgG (Invitrogen) was used as the secondary antibody. Photographic images were captured using an Olympus CX71 fluorescence microscope (Olympus).

The TUNEL assay was performed using a kit from Roche Applied Science (no. 1684795; Indianapolis, IN, USA) according to the manufacturer’s protocol. Briefly, 6-μm frozen tissue sections were rinsed three times in PBS and were incubated in 0.3% Triton X-100 (v/v) in 0.01 M PBS (pH 7.4) for 20 min at room temperature. Subsequently, the TUNEL reaction mixture was applied for 60 min at 37°C. The fluorescence signal was detected with an Olympus microscope (model CX71) at excitation/emission wavelengths of 492/520 nm.

Statistical analysis was performed using a one-tailed Student’s t-test (unilateral and unpaired). Kaplan-Meier survival plots were generated, and comparisons between survival curves were assessed using the log-rank statistical analysis. Data were analyzed using GraphPad Prism 5 software (San Diego, CA, USA), and a P-value <0.05 was considered to indicate a statistical significant difference.

In the present study, we collected 34 tissue specimens from gastric cancer patients and analyzed PDCD2 expression. The western blotting data showed that PDCD2 protein expression was significantly lower in the tumor tissues than that in the adjacent normal mucosa. We then analyzed expression of p53 protein and similar data were obtained (P<0.05; Fig. 1A). qRT-PCR data confirmed the western blotting data (P<0.05; Fig. 1B). To assess why p53 protein was more highly expressed in normal vs. tumor tissues, we performed immunohistochemistry and found that PDCD2 protein was localized in the cytoplasm, whereas p53 protein was localized in the nuclei of cells (Fig. 1C).

We then collected follow-up data from these patients and examined the association of PDCD2 expression with patient survival. Specifically, the 34 patients were followed up between 1 month and 5 years with a median time of 28 months. Kaplan-Meier analysis showed that expression of PDCD2 protein was associated with favorable prognosis, whereas p53 expression was associated with poor prognosis of the gastric cancer patients (P<0.05; Fig. 1D).

We next investigated the effects of exogenous PDCD2 expression in gastric cancer cell lines by transfecting PDCD2 cDNA into MKN28 and MKN45 cells. qRT-PCR and western blotting data confirmed the exogenous expression of PDCD2 in MKN28 and MKN45 cells after transfection (Fig. 2A and B). Similarly, we transfected p53 cDNA and shRNA into these two cell lines, respectively, and qRT-PCR and western blotting data confirmed p53 expression in MKN28 cells and p53 knockdown in MKN45 cells (Fig. 2C and D).

We then assessed phenotypic changes in these cells and found slower growth in the PDCD2-expressing MKN45 cells when compared with the growth of the untreated cells by using an [³H]thymidine incorporation assay (P<0.05; Fig. 3A). In addition, the cells showed a high level of apoptosis after PDCD2 cDNA transfection, as evidenced by Annexin V/PI double staining (P<0.05, Fig. 3B). PI staining of the cells revealed that PDCD2-expressing MKN45 cells were arrested in the S phase of the cell cycle (P<0.05, Fig. 3C), and CldU and IdU staining confirmed that the MKN45 cells were arrested at the early S phase (Fig. 3D). However, there was no effect of PDCD2 expression on the regulation of cell growth, apoptosis and cell cycle arrest in MKN45 cells transfected with p53 shRNA or parental cells (Fig. 3). Similarly, PDCD2 expression was able to inhibit proliferation, but induced apoptosis and early S phase arrest in p53-overexpressing MKN28 cells (Fig. 3). These results indicate that the antitumor activities of PDCD2 in gastric cancer cells are p53-dependent.

To determine whether p53 is required for the antitumor effects of PDCD2, we utilized p53^+/+ and p53^−/− nude mice and induced skin carcinogenesis with UVB irradiation. Expression and localization of p53 and PDCD2 proteins were assessed using immunofluorescence. The data revealed that PDCD2 protein was expressed in the skin tissues of both the p53^−/− and p53^+/+ nude mice after tail vein injection of pCDNA3.1-PDCD2 (Fig. 4A). In contrast, there was no p53 expression observed in the p53^−/− nude mice (Fig. 4A). After establishing UVB-induced skin tumorigenesis, PDCD2 expression resulted in a significant reduction in tumor number in the p53^+/+ nude mice. In contrast, there was no change in tumor number in the p53^−/− nude mice (P<0.05; Fig. 4B). Furthermore, we also found that PDCD2-transfected p53^+/+ mice had a prolonged survival rate when compared with the survival of the PDCD2-transfected p53^−/− mice (P<0.05; Fig. 4C). In addition, TUNEL staining of skin tissues showed that PDCD2 expression induced apoptosis in UVB-induced skin carcinoma of the p53^+/+ mice but not in the p53^−/− mice (Fig. 4D).

To explore the underlying mechanism responsible for PDCD2-induced apoptosis and cell cycle arrest in gastric cancer cells, we assessed the expression of cell cycle checkpoint proteins, including the protein kinase ATM and checkpoint kinase 1 and 2 (Chk1/Chk2), which are major components of the mechanisms that oversee the control of DNA replication and genomic integrity (10). We found that levels of Chk1 and Chk2 were not significantly changed in the MKN28 and MKN45 cells after PDCD2 transfection (Fig. 5). However, the levels of p-Chk1 and p-Chk2 were significantly increased in the MKN45 cells after PDCD2 transfection and in MKN28 cells after transfection with PDCD2 and p53 (Fig. 5). Furthermore, we found that the levels of ATM, the upstream protein of Chk1 and Chk2, were also increased (Fig. 5).