The human GC cell line HGC-27 was obtained from the Cell Line Bank, Chinese Academy of Sciences (Shanghai, China). The cell line was verified by polymerase chain reaction (PCR) and cultured without mycoplasma contamination; the species origin was also confirmed by PCR. In addition, the identity of the cell line was authenticated with short tandem repeat profiling. HGC-27 cells were cultured in RPMI-1640 (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% foetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc.) and incubated at 37°C in a humidified atmosphere containing 5% CO₂.

Cells were washed with ice-cold PBS and lysed in Tris-buffered saline (pH 7.4) containing 50 mmol/l Tris, 150 mmol/l NaCl, 1% Nonidet P-40, 1 mmol/l EDTA, 1 mmol/l Na₃VO₄, 10 mmol/l NaF, 2.5 mg/ml aprotinin and leupeptin, 1 mmol/l glycerophosphate plus 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride and 10 mmol/l iodoacetate, on ice for 15 min prior to removing cellular debris and nuclei by centrifugation at 10,000 × g for 5 min at 4°C. Cell lysates were incubated with the corresponding primary antibody CDK5 (cat. no. 2506; Cell Signaling Technology, Inc., Danvers, MA, USA) overnight at 4°C. Protein A-Sepharose beads (Amersham; GE Healthcare, Chicago, IL, USA) in a 50:50 mixture in 50 mmol/l Tris buffer, pH 7.0, were added and further incubated for 4 h at 4°C. The immunoprecipitates were washed 4 times with Tris-buffered saline and boiled for 5 min in 40 l of Laemmli buffer containing 0.02% blue bromophenol and 2% mercaptoethanol.

The cells were seeded into 60-mm dishes at a concentration of 5×10⁴ cells/well (20–30% confluence) and cultured to 80% confluence. The cells were then scraped and lysed in Radioimmunoprecipitation Assay buffer, and the lysates were centrifuged at 10,000 × g (4°C for 10 min). Protein concentrations were determined using the BCA Protein Assay kit (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. A total of 40 µg of protein from each sample was denatured, loaded into each well, separated by 10% SDS-PAGE, and transferred to a polyvinylidene difluoride membrane (EMD Millipore, Billerica, MA, USA). The membranes were blocked with 5% non-fat milk at room temperature for 1 h and then incubated overnight with primary antibodies in TBST (1L TBS buffer with 0.5 ml, 0.05% Tween-20; 1:1,000). Following washing with TBST, the membranes were incubated for 1 h at room temperature with the corresponding horseradish peroxidase (HRP)-conjugated secondary antibodies at appropriate dilutions and then washed three times with TBST. The proteins on the membranes were visualized using enhanced chemiluminescence (Amersham; GE Healthcare). Primary antibodies against enhanced green fluorescent protein (EGFP; cat. no. 2956; 1:1,000 dilution), CDK5/EGFP-CDK5 (cat. no. 2506; 1:1,000), PP2A/PP2Ac (cat. no. 2038; 1:1,000), Lamin A (cat. no. 86846S; 1:1,000 dilution), a-tubulin (cat. no. 2144; 1:1,000) and GAPDH (cat. no. 5174S; 1:1,000) were purchased from Cell Signaling Technology, Inc. HRP-conjugated goat anti-rabbit immunoglobulin (Ig)-G (cat. no. A4914) and anti-mouse IgG (cat. no. A0168) were purchased from Sigma-Aldrich; Merck KGaA (Darmstadt, Germany).

A nano-ultra performance (UP)-LC system (Waters Corporation, Milford, MA, USA) was used to separate the peptides. Samples were loaded on a trap column and flushed with mobile phase A (0.1% formic acid in H2O) at 5 µl/min for 3 min prior to being delivered onto a nanoUPLC column (C18, 150×0.075 mm, 1.7 µm). The peptides were eluted using a 7–45% B gradient (0.1% formic acid in acetonitrile) over 90 min into a nano-electrospray ionization LTQ Orbitrap mass spectrometer (Thermo Fisher Scientific, Inc.). The mass spectrometer was operated in data-dependent mode in which an initial FT scan recorded the mass range of 350–2,000 m/z, and the eight most abundant ions were automatically selected for collisional activated dissociation. The spray voltage was set as 2.0 kV. The normalized collision energy was set at 35% for MS/MS. Raw data were compared against the Uniprot human protein database (www.uniprot.org/uniprot/) containing 98,778 sequence entries using the SEQUEST algorithm embedded in the Protein Discoverer 1.3 Software (Thermo Fisher Scientific, Inc.). The following parameters were applied during the database search: 10 ppm precursor mass error tolerance, 1 Da fragment mass error tolerance, static modifications of carbamidomethylation for all cysteine residues, flexible modification of oxidation modifications for methionine residues, and one missed cleavage site of trypsin was allowed. False discovery rate <0.01 was used as filtering criteria for all identified peptides. Only proteins identified with ≥2 unique peptides were considered, and proteins identified with the same set of peptides were grouped.

EGFP-CDK5-nuclear localisation signal (NLS), EGFP-CDK5-nuclear export signal (NES), EGFP-CDK5-wild-type (WT), pTRE-EGFP, short hairpin (sh)-CDK5 (Lenti-shCKD5) and the corresponding empty vectors Lenti-scramble (Lenti-scr) were constructed, packed and purified by Shanghai GeneChem Co., Ltd. (Shanghai, China) and were transfected into the HGC-27 cell line (5×10⁴ cells/well in 6-well plates) with the Xfect Transfection Reagent (cat. nos. 631317 and 631318; Clontech Laboratories, Inc., Mountain View, CA, USA) according to the manufacturer's protocol. EGFP-CDK5-WT plasmids were then introduced using doxycycline (Sigma-Aldrich; Merck KGaA; cat. no. D9891). Then, once selected using puromycin or G418 (Sigma-Aldrich; Merck KGaA), sTable cells were harvested and stored in liquid nitrogen until subsequent use. CKD5 knockdown using Lenti-shCDK5, HGC-27 cells were transfected with shPP2A (lenti-shPP2A) and control plasmid (Shanghai GeneChem Co., Ltd.) according to the manufacturer's protocol using Lipofectamine™ 3000 Transfection Reagent (Thermo Fisher Scientific, Inc.). All of the sTable cells were then stored in liquid nitrogen for subsequent experimentation within 6 months post-transfection. CDK5-overexpressing GC cells were treated with Okadaic Acid (Cell Signaling Technology, Inc.; cat. no. 5934) at 100 nM for 30 min according to the manufacturer's protocol prior to subsequent experiments.

Cells were grown on glass coverslips, washed twice with PBS, fixed with PBS containing 4% formaldehyde at 4°C for 10 min, and permeabilized with 0.2% Triton X-100 in PBS at 4°C for 10 min. Following washing with PBS, cells were blocked with 10% goat serum (Abcam, Cambridge, MA, USA; cat. no. ab7481) at room temperature for 2 h. Then cells were incubated overnight at 4°C with primary antibody against CDK5 (Abcam; cat. no. ab151233; 1:1,000), washed with PBS and incubated with Alexa Fluor 488 IgG (Invitrogen; Thermo Fisher Scientific, Inc; cat. no. A-11008; 1:100; according to the manufacturer's instructions, cells were incubated with 4 µg/ml PBS containing 0.2% bovine serum albumin for 45 min at room temperature.). Alexa Fluor^® 568 phalloidin (Invitrogen; Thermo Fisher Scientific, Inc.) was employed to observe the lamellipodia, and nuclei were stained with DAPI (Sigma-Aldrich; Merck KGaA) at 37°C for 5 min. The coverslips were mounted with the SlowFade^® Gold reagent (Invitrogen; Thermo Fisher Scientific, Inc.) and observed under a laser confocal scanning microscope.

Cell migration (using non-Matrigel coated pore membrane) and invasion assays (using Matrigel coated pore membrane) were performed in Transwell chambers (polycarbonate filters of 8-mm porosity; BD Biosciences, Franklin Lakes, NJ, USA). Cells with serum-free RPMI-1640 were seeded in the upper chamber (5×10⁴ cells/well) for 24 h at 37°C, then medium in the lower chamber was replaced with complete growth medium containing 5% FBS (Gibco; Thermo Fisher Scientific, Inc.) and chambers were cultured for 45 h at 37°C. Non-invading cells were removed from the upper surfaces of the invasion membranes and the cells on the lower surface were stained with 0.1% crystal violet for 5 min at room temperature. Cells at the subcellular level of the microporous membrane were observed and photographed under an optical microscope (magnification, ×200).

Colony formation assays were performed to assess the ability of the cells to form colonies. Cell Counting Kit-8 (CCK-8; Dojindo Molecular Technologies, Inc., Kumamoto, Japan) and Sulforhodamine B (SRB; Sigma-Aldrich; Merck KGaA; cat. no. S9012) cell proliferation assays were performed to examine cell viability and cell proliferation. Briefly, cells were seeded into 96-well tissue culture plates at 1,000 cells/well. Plates were incubated for 24, 48, 72, 96 or 120 h at 37°C with 5% CO₂ in a humidified incubator. Once cells were treated and following incubation, 20 µl of CCK-8 was added to each well and cells were further incubated for an additional 3 h. Plates were read on a Biorad ELISA plate reader (Biotek Instruments, Inc., Winooski, VT, USA) using a 450-nm filter. Results of at least three independent experiments were analysed in duplicate. The relative cell proliferation ratios were plotted along with non-treated controls to determine the 100% level of activity. For the SRB colorimetric assay, cells were seeded into 96-well tissue culture plates at 1,000 cells/well and allowed to adhere for 5 days. Following 24, 48, 72, 96 or 120 h, cells were harvested and subjected to the SRB protocol, as described previously (32).

GC tissues and the respective adjacent non-tumour tissues were obtained following R0 gastrectomy procedures from 80 patients admitted to Fujian Medical University Union Hospital (Fujian, China) between February 2013 to February 2014 as well as detailed information regarding the clinicopathological parameters. All of the patients underwent R0 gastrectomy. None of the patients underwent preoperative chemotherapy or radiotherapy. Cases with the following characteristics were also excluded: Remnant GC, treated with palliative surgery and concurrent malignancy in other organs. Postoperative adjuvant chemotherapy was performed using 5-fluorouracil-based drugs plus oxaliplatin in advanced cases. The pathological stage of the tumour was reassessed according to the 2010 International Union Against Cancer Tumour-Node-Metastasis (TNM) classification on GC (7th edition) (33). The present study was approved by the Ethics Committee of Fujian Medical University Union Hospital (no. UH-2013009), and written informed consent was obtained from all patients involved.

RNA isolation was performed using TRIzol reagent (Thermo Fisher Scientific, Inc.) according to the manufacturer's manual. RT-PCR was performed with the ReverTra Ace qPCR RT kit (Toyobo Life Science, Osaka, Japan) using oligo-dT-primers and 1 µg of mRNA/reaction as a template. The temperature protocol for RT-PCR was as follows: 25°C for 5 min, 42°C for 60 min and 70°C for 5 min. RT-qPCR analysis was performed in triplicate on the ABI PRISM 7500 Sequence Detection System (Thermo Fisher Scientific, Inc.) using the Thunderbird SYBR qPCR Mix (Toyobo Life Science) according to the manufacturer's recommendations. The thermocycling conditions for qPCR were as follows: 40 cycles at 95°C for 3 min, 95°C for 15 sec and 60°C for 30 sec. The relative expression of PP2A was normalized to GADPH expression levels. The primer sequences were as follows: PP2A, 5′-CGAGAGTCATGACACGGAGG-3′ (forward) and 5′-GAGAACGCTTCCATTTGCCC-3′ (reverse); and GAPDH: 5′-GGTCGGAGTCAACGGATTTG-3′ (forward) and 5′-ATGAGCCCCAGCCTTCTCCAT-3′ (reverse). RNA expression levels were obtained using the comparative 2^−∆∆Cq method (34).

All patients were systematically followed up by trained doctors based on the institutional follow-up protocol using several approaches, including outpatient service, written letters, telephone calls, mail or visitation. Follow-up was conducted every 3 months during the first year and every 6 months from the second year onwards, and all surviving patients were followed for >3 years. The survival time was defined as the time from the date of surgery until the date of last contact or mortality. All 80 patients involved in the immunohistochemical analysis were followed up, and none were lost.

All statistical analyses were performed using SPSS 18.0 for Windows (SPSS, Inc., Chicago, IL, USA) and GraphPad Prism 5.0 software (GraphPad Software, Inc., La Jolla, CA, USA). Chi-squared tests were used to evaluate differences in proportions. Differences among groups were assessed by one-way analysis of variance with a Student-Newman-Keuls post hoc test, as previously described (35). Univariate survival analysis was performed using the Kaplan-Meier method, and the curves were compared using the log-rank test. Multivariate analysis was performed using the Cox proportional hazards model to further evaluate all significant prognostic factors that were identified in the univariate analysis. P<0.05 was considered to indicate a statistically significant different; all P-values were two-sided.

To examine the function of CDK5 in different subcellular locations, the present study first overexpressed EGFP-CDK5-WT, EGFP-CDK5-NLS, EGFP-CDK5-NES or p-TRE-EGFP in the HGC-27 GC cell line (Fig. 1A-C). Then, CCK-8 and SRB assays were performed, which revealed that cell proliferation and colony formation were significantly inhibited by nuclear CDK5 expression (Fig. 1D and E). Similarly, nuclear CDK5 expression markedly reduced the colony formation of HGC-27 GC cells and the metastasis of HGC-27 cells in Transwell assays. (Fig. 1F and G). Therefore, in vitro, CDK5 overexpression inhibited HGC-27 cell proliferation, colony formation and metastasis, which was consistent with our previous data obtained with the SGC-7901 and MGC-803 GC cell lines (26).

Furthermore, to examine the effect of CDK5 on GC tumourigenicity, HGC-27 cells with CDK5 stably knocked down and control cells were generated (Fig. 2A). In vitro, it was observed that no significant differences were evident in terms of proliferation and colony formation between the shCDK5 cells and control cells (Fig. 2B and C). However, the knockdown of CDK5 in HGC-27 GC cells promoted metastasis (Fig. 2D).

To investigate the functional mechanism of CDK5 in GC, the present study performed IP and LC-MS to identify potential targets of CDK5 in HGC-27 GC cells overexpressing EGFP-tagged CDK5-NLS, EGFP-tagged CDK5-NES or p-TRE-EGFP (Fig. 3A). Seven proteins were expressed in all of the groups (Fig. 3B). Further examination by co-IP and western blot analysis revealed that CDK5 interacted with PP2A (Fig. 3C). Therefore, it was hypothesized that CDK5 in combination with PP2A may serve a crucial role in GC.

Next, cells were treated with okadaic acid, an inhibitor of PP2A, and alterations in the proliferation, colony formation and metastasis of HGC-27 GC cells overexpressing CDK5 and in the control group were observed. No significant differences were identified between the two groups in terms of proliferation (Fig. 4A), though a reduction in the metastasis of GC cells overexpressing CDK5 was revealed (Fig. 4B). Furthermore, knockdown of PP2A reduced metastasis (Fig. 4C and D) in CKD5 knockdown (shCDK5 #1) HGC-27 cells, but it did not affect proliferation (Fig. 4E). Consequently, it was concluded that CDK5 may influence the metastasis of GC by interacting with PP2A.

In the present study, it was demonstrated that PP2A interacted with CDK5 and subsequently influenced the metastasis of GC. For further analysis, RT-qPCR was performed to examine the relative expression levels of PP2A in 80 pairs of GC tissues matched with normal tissues (Fig. 5A and B). PP2A expression in the GC tissues was lower than in the corresponding adjacent non-tumour tissues (Fig. 5).

To analyse the association between PP2A expression and clinicopathological features, specimens were collected from 80 patients with GC. The cohort was divided into groups: Low PP2A expression group and a high PP2A expression group, according to the median PP2A value. No significant differences were observed in terms of the clinical and pathological characteristics between the high- and low-PP2A groups (Table I). To investigate whether PP2A could serve as a prognostic marker, overall survival (OS) curves were plotted using the Kaplan-Meier method according to the PP2A expression level. Patients with low levels of PP2A had a poorer OS rate when compared with those with high levels of expression (P=0.035; Fig. 6). In addition, to assess whether the use of PP2A to predict OS was independent of the other clinical or pathological factors of the patients with GC, univariate and multivariate Cox proportional hazards analyses were performed. The results revealed that the PP2A levels and TNM stage were independent prognostic factors for OS in patients with GC (Table II). Therefore, low levels of PP2A may be indicative of a poor prognosis in patients with GC.