A role for p21-activated kinase 7 in the development of gastric cancer

Authors


Correspondence

Y. Liu, Department of General Surgery, Xinhua Hospital, Affiliated to School of Medicine, Shanghai Jiao Tong University, 1665 Kongjiang Road, Shanghai 200092, China

Fax/Tel: +86 21 65793206

E-mail: liuybphd@126.com

Abstract

p21-activated kinase (PAK)7 (also known as PAK5) is a member of the group B PAK family of serine/threonine protein kinases, which are effectors of the small GTPases Rac and CDC42. PAK7 can promote neurite outgrowth, induce microtubule stabilization, and activate cell survival signaling pathways. However, the role of PAK7 in cancer is still poorly understood. Here, we showed that PAK7 expression was upregulated in different gastric cancer cell lines and gastric cancer tissues, as compared with human embryonic kidney 293 cells and adjacent normal tissues, respectively. The results suggested that PAK7 expression was related to gastric cancer progression. Thus, we employed lentivirus-mediated small interfering RNA to inhibit PAK7 expression, to investigate the role of PAK7 in human gastric carcinogenesis. RNA interference efficiently downregulated expression of PAK7 in SGC-7901 and MGC-803 cells at both mRNA and protein levels. Knockdown of PAK7 inhibited human gastric cancer cell proliferation by inducing cell cycle arrest in G0/G1 phase, in concordance with the downregulation of CDK2, CDC25A, and cyclin D1. Our data suggest that PAK7 is a new hallmark of gastric cancer, in which PAK7 might contribute to gain of tumor growth potential, acting by affecting the expression of cell cycle regulators. Therefore, PAK7 may be an attractive candidate as a therapeutic target in gastric cancer.

Abbreviations
CDK

cyclin-dependent kinase

CRC

colorectal carcinoma

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

GFP

green fluorescent protein

HEK293

human embryonic kidney 293

IHC

immunohistochemistry

MOI

multiplicity of infection

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide

PAK

p21-activated kinase

PI

propidium iodide

RNA si-CTRL

SGC-7901 and MGC-803 cells transduced with scrambled siRNA

SD

standard deviation

shRNA

short-hairpin

si-PAK7

SGC-7901 and MGC-803 cells transduced with PAK7 siRNA

siRNA

small interfering RNA

Introduction

The incidence of gastric cancer has gradually decreased in recent years, but it still remains the second leading cause of cancer deaths worldwide [1]. Surgical resection is the mainstay of treatment, and can cure patients with early-stage cancer. However, the majority of patients suffering from gastric cancer are diagnosed at an advanced stage [2, 3]. Although considerable improvements have been made in chemotherapy, radiotherapy, and surgical techniques, the survival rate of patients with advanced gastric cancer is low [4]. The development and progression of gastric cancer have now been shown to involve a number of genetic alterations of tumor suppressor genes and/or oncogenes [5]. Thus, it is of great clinical value to further understand the molecular mechanisms involved in gastric cancer and to find valuable diagnostic markers and novel therapeutic targets.

The p21-activated protein kinases (PAKs) are a family of serine/threonine protein kinases that are known to be effectors of the small GTPases Rac and CDC42 [6, 7]. PAKs are activated in response to extracellular signals, and regulate cell morphology, cell motility, and cell survival [8-10]. PAKs have been shown to have various roles in cancer initiation and progression [9, 11, 12]. The six members of the PAK family are subdivided into two groups: group A contains PAK1–3, and group B contains PAK4–6 [13]. Among the three group B PAKs, PAK7 (also known as PAK5) is the least understood. It is predominantly expressed in the brain, and promotes neurite outgrowth in mouse neuroblastoma N1E-115 cells by downregulation of RhoA activity [14]. PAK7 is autophosphorylated, and the GTPases Rac and CDC42 do not regulate PAK7 activity. However, CDC42, but not Rac, is able to activate the autophosphorylation of PAK7 in a GTP-dependent manner [15]. PAK7 is an inhibitor of MARK2, a kinase that has been shown to induce microtubule disruption by phosphorylating microtubule-associated proteins such as tau. A direct interaction between PAK7 and MARK2 leads to stable microtubules and dynamic actin [16, 17]. PAK7 has different effectors, depending on its localization. In the cytosol, it can activate the c-Jun N-terminal kinase pathway, which has been implicated in the regulation of cell growth in mammalian cells [14, 18]. In mitochondria, it can activate cell survival signaling pathways by phosphorylating BAD on Ser112 and preventing its translocation to the mitochondria [19, 20]. It has been reported that the mitochondrial localization of PAK7 is independent of its kinase activity or its ability to bind to CDC42 [20].

However, the role of PAK7 in cancer is still poorly understood, apart from its antiapoptotic properties [19]. It has been demonstrated that point-mutated PAK7 contributes to human cancer [21], and that PAK7 is involved in the malignant progression of colorectal carcinoma (CRC) [22, 23]. In the present study, we found that PAK7 was overexpressed in human gastric cancer tissues and cell lines. We showed that specific downregulation of PAK7 inhibited SGC-7901 and MGC-803 cell proliferation by inducing cell cycle arrest in G0/G1 phase, possibly acting by affecting the expression of cell cycle regulators. Our work is the first to demonstrate that PAK7 plays a significant role in human gastric cancer pathogenesis and progression. Thus, PAK7 may be an attractive candidate as a therapeutic target in gastric cancer.

Results

Immunohistochemical determination of PAK7 expression in human gastric cancer tissues

To determine the potential role of PAK7 in human gastric cancer progression, we evaluated PAK7 expression in surgical specimens of human gastric cancer by immunohistochemistry (IHC). Representative immunohistochemical stains are shown in Fig. 1, and the entire dataset is shown in Table 1. Among the 57 specimens of gastric cancer tissue that were stained by IHC, five (8.8%) samples were classified into the strong expression group (+++, hadro-positive), 18 (31.6%) into the moderate expression group (++, positive), and 21 (36.8%) into the weak expression group (+, weakly positive); only 13 (22.8%) showed negative staining. Thus, the positive ratio was 77.2%. In contrast, among the 57 specimens of adjacent normal tissues, one (1.8%) sample was classified into the strong expression group, and seven (12.3%) into the moderate expression group, 19 (33.3%) into the weak expression group; 30 (52.6%) showed negative staining. Thus, the positive ratio was 47.4%. Immunohistochemical analysis demonstrated that PAK7 expression was upregulated significantly in gastric cancer tissues, as compared with adjacent normal tissues (P = 0.0001). We also determined the relationship between the immunohistochemical expression level of PAK7 and clinicopathological parameters of gastric cancer patients. However, the results demonstrated that none of the clinicopathological parameters was affected by PAK7 expression (Table 2).

Table 1. Expression pattern of PAK7 in human gastric cancer tissues and adjacent tissues
Type of tissueNumber of casesPAK7 expressionP-value
Negative (–)Weakly positive (+)Positive (++)Hadro-positive (+++)
  1. a

    χ2-test.

Cancer tissues5713211850.0001a
Adjacent tissues57301971
Table 2. Relationship between clinicopathological parameters and PAK7 immunohistochemical expression
Clinicopathological parameters N Cancer tissuesAdjacent tissues χ 2 P-value
Positive, no. of cases (%)Negative, no. of cases (%)Positive, no. of cases (%)Negative, no. of cases (%)
Age (years)2.1760.537
< 602218 (81.8)4 (18.2)8 (36.4)14 (63.6)  
≥ 603526 (74.3)9 (25.7)19 (54.3)16 (45.7)
Gender3.8350.280
Male4435 (79.5)9 (20.5)18 (40.9)26 (59.1)  
Female139 (69.2)4 (30.8)9 (69.2)4 (30.8)
Tumor size (cm)1.4280.670
< 53324 (72.7)9 (27.3)17 (51.5)16 (48.5)  
≥ 52420 (83.3)4 (16.7)10 (41.7)14 (58.3)
Histology3.3360.993
Tubular adenocarcinoma2922 (75.9)7 (24.1)15 (51.7)14 (48.3)  
Mucinous adenocarcinoma11 (100.0)0 (0.0)1 (100.0)0 (0.0)
Signet ring cell carcinoma22 (100.0)0 (0.0)1 (50.0)1 (50.0)
Poorly differentiated adenocarcinoma2217 (77.3)5 (22.7)10 (45.5)12 (54.5)
Undifferentiated carcinoma33 (100.0)0 (0.0)1 (33.3)2 (66.7)
Infiltration0.1230.989
T1 and T21612 (75.0)4 (25.0)8 (50.0)8 (50.0)  
T3 and T44132 (78.0)9 (22.0)19 (46.3)22 (53.7)
Histology differentiation0.7850.853
Well or moderate2921 (72.4)8 (27.6)14 (48.3)15 (51.7)  
Low2823 (82.1)5 (17.9)13 (46.4)15 (53.6)
TNM stage1.5080.680
I + II2316 (69.6)7 (30.4)10 (43.5)13 (56.5)  
III + IV3428 (82.4)6 (17.6)17 (50.0)17 (50.0)
Venous invasion0.01980.999
No3628 (77.8)8 (22.2)17 (47.2)19 (52.8)  
Yes2116 (76.2)5 (23.8)10 (47.6)11 (52.4)
Lymphatic invasion0.9730.808
No1712 (70.6)5 (29.4)7 (41.2)10 (58.8)  
Yes4032 (80.0)8 (20.0)20 (50.0)20 (50.0)
Distant metastasis1.4320.698
No5643 (76.8)13 (23.2)26 (46.4)30 (53.6)  
Yes11 (100.0)0 (0.0)1 (100.0)0 (0.0)
Figure 1.

Immunohistochemical analysis of PAK7 expression in human gastric cancer and adjacent normal tissue. (A–C) Weakly positive (A), positive (B) and hadro-positive (C) PAK7 staining in the gastric cancer tissue. (D–F) Negative (D), weakly positive (E) and positive (F) PAK7 staining in the adjacent normal tissue; Original magnification: × 20 objective lens.

Western blotting analysis of PAK7 expression in human gastric cancer cell lines

We also examined PAK7 expression in human gastric cancer cell lines. We used three gastric cancer cell lines (MKN-45, SGC-7901, and MGC-803) and a human normal cell line, human embryonic kidney 293 (HEK293). As shown in Fig. 2A, we found that PAK7 expression was low in HEK293 cells, whereas all three gastric cancer cell lines showed high PAK7 expression. SGC-7901 and MGC-803 cells, in particular, expressed higher levels of PAK7 than MKN-45 cells, indicating that the SGC-7901 and MGC-803 cell lines are two suitable cell models for RNA interference-mediated disruption of PAK7 expression.

Figure 2.

Lentivirus-mediated siRNA decreased PAK7 expression in SGC-7901 and MGC-803 cells. (A) PAK7 protein was highly expressed in human gastric cancer cell lines (MKN-45, SGC-7901, and MGC-803) as shown by western blotting. HEK293 cells served as a negative control. (B) Lentivirus transduction efficiency was estimated 3 days after infection at an MOI of 40. GFP expression in transfected cells was observed under a light microscope and a fluorescence microscope. Upper: light micrograph. Lower: fluorescence micrograph. Magnification: × 100. (C) Total RNA was extracted at 5 days after infection, and relative PAK7 mRNA expression was determined by quantitative real-time RT-PCR. β-Actin was used as an internal control. Data represent the mean ± SD of three independent experiments. *P < 0.05 and **P < 0.01, as compared with si-CTRL. (D) Total cellular proteins were extracted at 5 days after infection, and determined by western blotting analysis with antibodies against PAK7 and GAPDH as an internal control. Data represent one of three separate experiments.

Lentivirus-mediated small interfering RNA (siRNA) effectively and particularly inhibited PAK7 mRNA and protein expression in SGC-7901 and MGC-803 cells

To determine the lentiviral transduction efficiency in SGC-7901 and MGC-803 cells, green fluorescent protein (GFP) expression was examined by microscopy at a multiplicity of infection (MOI) of 40 on day 3 after infection (Fig. 2B), and the efficiency of lentiviral transduction in SGC-7901 and MGC-803 cells was > 90%.

The mRNA and protein expression levels of PAK7, inhibited by lentivirus-mediated siRNA in SGC-7901 and MGC-803 cells, were analyzed by real-time PCR and western blot. As compared with SGC-7901 and MGC-803 cells transduced with scrambled siRNA (si-CTRL) as negative controls, the levels of PAK7 mRNA in SGC-7901 and MGC-803 cells transduced with PAK7 siRNA (si-PAK7) were significantly decreased, by 74% (P < 0.01) and 68% (P < 0.01), respectively (Fig. 2C). Additionally, western blot also showed that the PAK7 protein expression levels in SGC-7901 and MGC-803 cells were significantly reduced in the si-PAK7 group as compared with the si-CTRL group (Fig. 2D).

These results indicated that lentivirus-mediated siRNA effectively and specifically suppressed PAK7 expression in SGC-7901 and MGC-803 cells.

Effect of PAK7 downregulation on the proliferation of SGC-7901 and MGC-803 cells

To further evaluate whether PAK7 gene silencing in SGC-7901 and MGC-803 cells may inhibit cell growth and proliferation, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) and colony formation assays were performed. As shown in Fig. 3A,B, the cell viability of the transfected cells was detected by MTT assay daily for 5 days, and downregulation of PAK7 expression decreased the viability of SGC-7901 and MGC-803 cells in a time-dependent manner. After 5 days of infection, the percentage of viable cells in the si-PAK7 group decreased by ~ 29.7% and 31.2%, respectively, as compared with the si-CTRL group (P < 0.01 and P < 0.01, respectively).

Figure 3.

Lentivirus-mediated siRNA targeting PAK7 inhibited SGC-7901 and MGC-803 cell proliferation. (A) Cell proliferation of untransfected or transfected SGC-7901 cells was measured by MTT assay once daily for 5 days. (B) Cell proliferation of untransfected or transfected MGC-803 cells was measured by MTT assay once daily for 5 days. Cell proliferation was expressed as the absorbance values. (C–F) SGC-7901 cells (C, D) and MGC-803 cells (E, F) were seeded at 200 per well and allowed to form colonies. Colony numbers were counted and recorded. The colonies stained with Giemsa were observed under a light microscope.

Furthermore, the results of the colony formation assay also showed that lentivirus-mediated siRNA against PAK7 caused significant decreases (55.2% and 42.8%, respectively) in the number of colonies in SGC-7901 and MGC-803 cells, as compared with the si-CTRL group (P < 0.01 and P < 0.05, respectively; Fig. 3C–F).

These results suggested a significant role for PAK7 in gastric cancer cell growth and proliferation.

Downregulation of PAK7 expression induced cell cycle arrest at G0/G1 phase by affecting the expression of cell cycle regulators in SGC-7901 and MGC-803 cells

To explore the possible underlying mechanisms of PAK7 suppression in inhibiting SGC-7901 and MGC-803 cell proliferation, the effect of PAK7 silencing on cell cycle progression was analyzed by flow cytometry. A statistical analysis was performed on the cell cycle findings. As shown in Fig. 4A,B, knockdown of PAK7 in SGC-7901 cells induced an increase in the percentage of cells in G0/G1 phase of the cell cycle (from 46.0% in the si-CTRL group to 55.2% in the si-PAK7 group, P < 0.001), parallel with a decrease in the percentage of cells in S phase (from 35.6% in the si-CTRL group to 28.8% in the si-PAK7 group, P < 0.01). Similarly, knockdown of PAK7 expression induced an increase in the percentage of MGC-803 cells at G0/G1 phase (Fig. 4C,D, P < 0.01). These results indicated that PAK7 knockdown could inhibit SGC-7901 and MGC-803 cell proliferation by the induction of G0/G1 arrest.

Figure 4.

Flow cytometric analysis of cell cycle distribution in SGC-7901 cells (A,C) and MGC-803 cells (B,D). Knockdown of PAK7 by RNA interference in SGC-7901 and MGC-803 cells induced cell cycle arrest in G0/G1 phase at 5 days after lentivirus transduction. Data represent the mean ± SD of three independent experiments. *P < 0.05 and **P < 0.01 as compared with si-CTRL.

Furthermore, the effect of PAK7 knockdown on the expression of cell cycle regulators was examined with real-time PCR and western blot. As shown in Fig. 5A,B, the protein and mRNA expression levels of CDK2, CDC25A and cyclin D1 in the si-PAK7 group were lower than those in the si-CTRL group. These results showed that knockdown of PAK7 induced cell cycle arrest in G0/G1 phase, possibly by affecting the expression of cell cycle regulators.

Figure 5.

Effect of PAK7 knockdown on cell cycle regulators. (A) Total RNA was extracted at 5 days after infection, and relative CDK2, CDC25A and cyclin D1 mRNA expression levels were determined by quantitative real-time RT-PCR. β-Actin was used as an internal control. Data represent the mean ± SD of three independent experiments. *P < 0.05 and **P < 0.01 as compared with si-CTRL. (B) Total cellular proteins were extracted at 5 days after infection, and determined by western blotting analysis with antibodies against CDK2, CDC25A, cyclin D1 and GAPDH as an internal control. Data represent one of three separate experiments.

Discussion

PAK7 can promote neurite outgrowth, induce microtubule stabilization, and activate cell survival signaling pathways [14-17]. However, little is known about the role of PAK7 in cancer. Recently, it has been reported that PAK7 is involved in the malignant progression of CRC, and may serve as a novel therapeutic target in CRC. PAK7 expression increases significantly during CRC progression, and PAK7 promotes CRC metastasis by regulating CRC cell adhesion and migration [22]. Accordingly, knockdown of PAK7 in LoVo cells resulted in increased apoptosis. Mechanistically, PAK7 directly phosphorylated BAD on Ser112, and indirectly led to phosphorylation of Ser136 via the Akt pathway in CRC cells [23]. Nevertheless, PAK7 has never been linked to gastric cancer to date.

In the present study, we showed, for the first time, that a significant increase in PAK7 expression was detected in the vast majority of human gastric cancer tissues as compared with adjacent normal tissues (Fig. 1; Table 1). PAK7 protein expression was also markedly increased in three human gastric cancer cell lines (MKN-45, SGC-7901, and MGC-803), particularly in SGC-7901 and MGC-803 cells, as compared with HEK293 cells (Fig. 2A). Therefore, the SGC-7901 and MGC-803 cell lines were chosen in our study as the cell line models with which to investigate the role of PAK7 in gastric cancer. These results also suggested that PAK7 expression was related to gastric cancer progression and could be a potential therapeutic target for gastric cancer treatment.

To further evaluate the biological significance of PAK7 in the pathogenesis of gastric cancer, we employed lentivirus-mediated siRNA to effectively and specifically silence endogenous PAK7 expression in the human gastric cancer cell lines SGC-7901 and MGC-803 (Fig. 2). Tumorigenesis, in general, is related to an imbalance between cell proliferation and cell death favoring growth [24]. We first examined whether knockdown of PAK7 would influence the proliferation of SGC-7901 and MGC-803 cells. MTT and colony formation assays indicated that knockdown of PAK7 significantly inhibited SGC-7901 and MGC-803 cell proliferation (Fig. 3). In addition, the decrease in cell viability can be partly attributed to the occurrence of G0/G1 cell cycle arrest after PAK7 silencing in SGC-7901 and MGC-803 cells, suggesting that PAK7 is involved in cell cycle regulation (Fig. 4).

The mammalian cell cycle is controlled by a series of highly regulated processes, and its dysregulation is a hallmark of human cancer [25, 26]. Progression through the cell cycle depends on the activation of cyclin-dependent kinases (CDKs) and their regulatory subunits, the cyclins [27]. Passage from G1 into S phase requires the activation of cyclin D1-associated CDK4 and CDK6 [28, 29] in addition to cyclin E/CDK2 [30], both of which contribute to phosphorylation of the retinoblastoma protein, pRb [31, 32]. Active cyclin–CDK complexes phosphorylate members of the pocket protein family (Rb, p107, and p130) to release and activate the E2F transcription factor and induce cell cycle progression. Cyclin–CDK complexes are inhibited by phosphorylation at two sites (Tyr14 and Tyr15 in CDK2) that are dephosphorylated by members of the CDC25 phosphatase family. Three CDC25 isoforms, A, B, and C, have been identified in mammalian cells. CDC25B and CDC25C play crucial roles at the G2/M transition. CDC25A mainly activates CDK2, and induces progression from G1 to S phase, which is rate-limiting for entry into S phase [33]. In the present study, we observed that knockdown of PAK7 inhibited human gastric cancer cell proliferation by inducing cell cycle arrest in G0/G1 phase, in concordance with the downregulation of CDK2, CDC25A and cyclin D1 at both the mRNA and protein levels (Fig. 5). Thus, PAK7 possibly acted by affecting the expression of cell cycle regulators. However, further studies are needed to elucidate the precise molecular mechanism by which PAK7 silencing induces cell cycle arrest in SGC-7901 and MGC-803 cells.

In conclusion, this study is the first to have examined the preliminarily mechanistic role of PAK7 in gastric cancer. Our results showed that lentivirus-mediated siRNA effectively and specifically suppressed PAK7 expression in SGC-7901 and MGC-803 cells. Knockdown of PAK7 significantly inhibited SGC-7901 and MGC-803 cell proliferation by inducing cell cycle arrest in G0/G1 phase, possibly acting by affecting the expression of cell cycle regulators. Thus, PAK7 is expected to be a potential therapeutic target for the treatment of gastric cancer.

Experimental procedures

Reagents and antibodies

RPMI-1640, fetal bovine serum and the bicinchoninic acid protein assay were from HyClone (Logan, USA); TRIzol Reagent and Lipofectamine 2000 were from Invitrogen (Carlsbad, CA, USA); the immunohistochemistry staining kit was from Peninsula Laboratories (San Carlos, CA, USA); MTT was from Dingguo Biology (Shanghai, China); dimethylsulfoxide was from Sibas (Shanghai, China); Giemsa was from Chemicon International (Temecula, CA, USA); propidium iodide (PI) was from Sigma-Aldrich (St Louis, MO, USA); Moloney murine leukemia virus reverse transcriptase was from Promega (Madison, WI, USA); oligo-dT was from Sangon Biotech (Shanghai, China); SYBR Green Master Mixture was from Takara (Otsu, Japan); pFUGW vector and virion-packaging elements (pVSVG-I and pCMVΔR8.92) were from Hollybio (Shanghai, China); mouse anti-PAK7, anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH), anti-CDK2, anti-CDC25A and anti-cyclin D1 IgG, and goat anti-(mouse IgG), were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). All other chemicals were of analytical grade.

Tissue specimens

IHC was performed for 29 cases of tubular adenocarcinoma, one case of mucinous adenocarcinoma, two cases of signet ring cell carcinoma, 22 cases of poorly differentiated adenocarcinoma, and three cases of undifferentiated carcinoma. These patients (N = 57) underwent surgery for gastric cancer at the Cancer Center at Xinhua Hospital, Affiliated to School of Medicine, Shanghai Jiao Tong University. The basic demographics of this group and the pathological characteristics are shown in Table 2. The samples were used with the written informed consent of the patient and the approval of the ethics committee of the Xinhua Hospital, Affiliated to School of Medicine, Shanghai Jiao Tong University.

IHC

For IHC, a labeled streptavidin–biotin method was used, with the Immunohistochemistry Staining Kit from Peninsula Laboratories, as described previously [34]. Formalin-fixed and paraffin-embedded tissue sections (5 μm in thickness) were deparaffinized with xylene, and hydrated in a graded ethanol series. In order to unmask antigen, slides were brought to the boil in 10 mm sodium citrate buffer (pH 6.0), and then maintained at a sub-boiling temperature for 10 min. Slides were cooled on the benchtop for 30 min. After being rinsed for 5 min in tap water, the tissue sections were placed in 3% H2O2 in distilled water for 30 min to block endogenous peroxidase activity. After being rinsed with NaCl/Pi, the slides were incubated with normal goat serum blocking solution for 30 min to block nonspecific binding, and then incubated with the primary antibody for 60 min at room temperature or overnight at 4 °C. After being again rinsed with NaCl/Pi, the slides were incubated with biotinylated secondary antibody for 30 min at room temperature, and then incubated with the streptavidin–horseradish peroxidaseconjugate for 30 min. A positive reaction was visualized by incubating the slides with stable diaminobenzidine and counterstaining with Mayer's hematoxylin for 2 min. After being rinsed with distilled water and dehydrated, the slides were mounted with mounting solution. Serum blocking solution in place of the primary antibody was used as a negative control.

Evaluation of PAK7 expression

The slides were evaluated in a double-blind manner by three pathologists independently, and the scores were the proportion of positive tumor cells and the intensity of the coloring, as described previously [35]. The results for the tissues were determined from at least 1000 cells that were counted systematically at × 400 magnification in 10 visual fields selected at random. The proportion of positive tumor cells was determined according to the following classification: 0, no cells stained; 1, more than one-third of cells stained; 2, one-third to two-thirds of cells stained; and 3, more than two-thirds of cells stained. The groups could also be classified into the following four groups by the intensity of the coloring: 0, no coloring; 1, stramineous; 2, buffy; and 3, dark brown. The two scores were combined to obtain the final score: 0, negative; 2–3, weakly positive (+); 4, positive (++); and 5–6, hadro-positive (+++).

siRNA design and construction of recombinant lentiviral vector

A third-generation self-inactivating lentivirus vector containing a cytomegalovirus-driven GFP reporter and a U6 promoter upstream of cloning restriction sites (NheI/PacI) to allow the introduction of oligonucleotides encoding short-hairpin RNA (shRNA) were described previously [36]. The cDNA sequence of PAK7 was obtained from GenBank (accession number: NM_020341). The shRNA target sequence for PAK7 was 5′-CGGGATTACCAC CATGACAAT-3′, which was subjected to blast analysis against the human genome database to eliminate cross-silence phenomena with nontarget genes. DNA oligonucleotides to produce plasmid-based shRNA were cloned into the pFUGW vector by use of NheI/PacI restriction sites. The lentiviral expression vector pFUGW carrying a scrambled fragment (5′-TTCTCCGAACGTGTCACGT-3′) that has no significant homology with mouse or human gene sequences was used as a negative control. The correct insertion of the specific shRNA was verified with restriction digestion analysis and plasmid DNA sequencing.

The lentiviral expression vector (pFUGW vector) and virion-packaging elements (pVSVG-I and pCMVΔR8.92) were cotransfected into 293T cells with Lipofectamine 2000, according to the manufacturer's instructions, for the generation of recombinant lentivirus. The supernatant was collected 48 h later, centrifuged (4000 g, 4 °C, 10 min) to remove cell debris, and then filtered through 0.45-μm cellulose acetate filters. The collected supernatant was concentrated again by spinning at 4000 g for 15 min and then at < 1 000 g for 2 min). The concentrated virus was stored at −80 °C. The lentiviral vectors expressed GFP, which allowed for titering and measurement of their infection efficiency in transfected cells. The viral titer was determined by counting GFP-positive cells after transfection.

Cell culture and lentiviral vector transduction in SGC-7901 and MGC-803 cells

The human gastric cancer cell lines MKN-45, SGC-7901 and MGC-803 were purchased from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). Cells were cultured in RPMI-1640 with 10% heat-inactivated fetal bovine serum, penicillin (100 U·mL−1), and streptomycin (100 U·mL−1). Cells were incubated at 37 °C in a humidified atmosphere of 95% air and 5% CO2. For SGC-7901 and MGC-803 cell infection, cells were subcultured at 5 × 104 cells per well into six-well culture plates. When they had grown to 30% confluence, cells were transduced with lentivirus-mediated PAK7 siRNA or scrambled siRNA at an MOI of 40. Cells were harvested 3 days after infection, and analyzed by fluorescence microscopy.

Quantitative real-time RT-PCR analysis

Total RNA was extracted with TRIzol Reagent, according to the manufacturer's instructions. RNase-free DNase I was used to remove DNA contamination. The quantity and purity of the RNA were determined by UV absorbance spectroscopy. The first-strand cDNA synthesis was performed with 2 μg of total RNA and Moloney murine leukemia virus reverse transcriptase, according to the manufacturer's instructions. The resulting cDNAs were subjected to real-time PCR analysis to evaluate the relative expression levels of PAK7, CDK2, CDC25A, cyclin D1, and β-actin (an internal control), with the following primers: 5′-ACAAGCACCAGCGAAAAGTT-3′ (forward) and 5′-TTTTCTTCCCAAA CATGATGC-3′ (reverse) for PAK7; 5′-CTGGACACT GAGACTGAGG-3′ (forward) and 5′-GAGGACCCGATGAGAATGG-3′ (reverse) for CDK2; 5′-ACACAGCAACTAGCCATCTCCAG-3′ (forward) and 5′-GCCAGCCTCCT TACCATCACG-3′ (reverse) for CDC25A; 5′-GGTGGCAAGAGTGTGGAG-3′ (forward) and 5′-CCTGGAAGTCAACGGTAGC-3′ (reverse) for cyclin D1; and 5′-GGCGGCACCACCATGTACCCT-3′ (forward) and 5′-AGGGGCCGGACTCGTCATACT-3′ (reverse) for β-actin. Real-time PCR reactions with SYBR Green Master Mixture were run on a TAKARA TP800-Thermal Cycler Dice Real-Time System. The cycling parameters were 95 °C for 15 s, 45 cycles of 95 °C for 5 s and 60 °C for 30 s, and a melting curve analysis. Ct was measured during the exponential amplification phase, and the amplification plots were analyzed with thermal dice real time system software Version 3.0 (TAKARA, Otsu, Japan). The relative expression levels of the target gene were normalized to that of the internal control gene, β-actin. The data were analyzed with the comparative threshold cycle (2−ΔΔCt) method [37].

Western blotting analysis

The transfected cells were washed twice with NaCl/Pi, and suspended in a lysis buffer (2% mercaptoethanol, 20% glycerol, and 4% SDS, in 100 mm Tris/HCl buffer, pH 6.8). After 15 min of incubation on ice, the cells were disrupted by ultrasound on ice. The lysates were cleared by centrifugation (12 000 g) at 4 °C for 15 min. The protein concentration was determined with a bicinchoninic acid protein assay. Equal amounts of protein were subjected to 8% SDS/PAGE, and transferred to poly(vinylidene difluoride) membranes. After blocking of the nonspecific binding sites for 1 h with 5% nonfat milk in NaCl/Tris-Tween at room temperature, the membranes were incubated with primary antibodies overnight at 4 °C. Then, the membranes were subjected to three 15-min washes with NaCl/Tris-T, and incubated with horseradish peroxidase-conjugated secondary antibody for 2 h at room temperature. After being washed three times with NaCl/Tris-T, detected proteins were visualized on an ECL plus western blotting detection system (Amersham Biosciences, Little Chalfont, UK). GAPDH protein levels were used as a control to verify equal protein loading.

Cell proliferation assay

Cell proliferation was assessed with the MTT assay as described by Mossman [38]. The MTT assay is based on the conversion of the yellow tetrazolium salt MTT to purple formazan crystals by metabolically active cells. It provides an estimate of the number of viable cells. In brief, the transfected cells were seeded in 96-well plates at 2 × 103 cells per 100 μL per well, and cultured in a 37 °C, 5% CO2 incubator. Ten microliters of MTT solution (5 mg·mL−1) was added to each well once daily for 5 days, and plates were incubated for 4 h at 37 °C. After removal of the supernatant, 100 μL of dimethylsulfoxide was added to dissolve the crystals. The absorbance at 490 nm was detected with a microplate reader (Bio-Rad 680, Hercules, CA, USA). The growth curve was constructed according to the absorbance values at 490 nm.

Colony formation assay

Both nontransfected and transfected SGC-7901 and MGC-803 cells (200 cells per well) were seeded in six-well plates. The culture medium was changed at regular time intervals. After 14 days of culture, adherent cells were washed twice with NaCl/Pi, and fixed with 4% paraformaldehyde for 30 min at room temperature. The colonies were stained with Giemsa solution for 15 min, and then washed with water and air-dried. Cell colonies were counted with a light microscope. The experiment was performed in triplicate.

Analysis of cell cycle distribution

The effect of PAK7 knockdown on cell cycle progression was determined by flow cytometry following staining with PI, as described previously [39]. In brief, the transfected cells were seeded onto six-well plates (1 × 105 per well). Five days after lentivirus transduction, cells were harvested, fixed in 70% ethanol, and stored overnight at 4 °C. For analysis, 1 mL of freshly prepared NaCl/Pi staining solution (50 μg·mL−1 PI and 100 μg·mL−1 RNase A) was added to the cells. Following incubation for 1 h in the dark at room temperature, cells were analyzed by flow cytometry (FACSCalibur; Becton Dickinson, San Jose, CA, USA). The fractions of the cells in G0/G1, S and G2/M phases were analyzed with dedicated software (Becton Dickinson).

Statistical analysis

Statistical analysis was performed with spss version 10.0 (SPSS, Chicago, IL, USA). Data were expressed as the mean of at least three different experiments ± standard deviation (SD). The statistical difference between PAK7 expression in human gastric cancer tissue and adjacent normal tissues was evaluated in a cross-table with the chi-square-test. Other results were analyzed with Student's t-test, and statistically significant differences were defined as P < 0.05.

Acknowledgements

This study was supported by the Research Foundation of Shanghai Xinhua Hospital (No. 12YJ01) and Basic Research Foundation of Science and Technology Committee of Shanghai (No. 12JC1406700).

Ancillary