This article is mentioned in:

Introduction

Despite recent advances in surgical treatment, chemotherapy and radiotherapy, gastric cancer remains a major global health burden. The most recent estimates found that gastric cancer is the fourth most common cancer and the second most common cause of cancer mortality worldwide (1,2). The majority of gastric cancer mortalities are a result of complications associated with recurrence and metastasis, even following curative resection. However, the invasion and metastasis of gastric cancer is a complex multistep process, in which an invasive capacity is required for gastric cancer cells to undergo metastasis. The aberrant expression of specific genes is known to contribute to this transformation (3).

T-cell lymphoma invasion and metastasis inducing factor 1 (Tiam-1), an important member of the Dbl oncogene family, was identified in T-lymphoma cells using proviral tagging in combination with in vitro selection for invasiveness (4). The human Tiam-1 gene maps to the syntenic region (q22) on human chromosome 21 (21q22) and encodes a 170 kDa transmembrane glycoprotein with intrinsic Rho GTPase activity. The gene contains a Dbl homologous (DH) domain adjacent to a pleckstrin homologous domain, a typical structure of guanine nucleotide exchange factors (GEFs) (5). As a specific GEF for Ras-related C3 botulinum toxin substrate 1 (Rac-1), a member of the Rho oncogene family, Tiam-l catalyzes the transition of Rac-l from an inactive GDP-bound state to an active GTP-bound state. In its active GTP-bound state, Rac-1 activates downstream signaling pathways associated with a number of important cellular events, including cytoskeletal reorganization, cell adhesion and migration (6). In addition, Tiam-1 has been demonstrated to increase the invasion of T-lymphoma cells, stimulate the cellular migration of fibroblasts and promote the motility of specific neurocytes. An increasing number of studies are focusing on Tiam-1 regulatory mechanisms, as well as its role in tumor progression and metastasis. Increased Tiam-1 expression has been revealed to correlate with breast cancer grade in humans and the metastatic potential of human breast cancer cell lines in nude mice (7), moreover Tiam1 may represent a marker of progression and metastasis of hepatocellular carcinomas and colon tumors (8–10). In our previous study, negative staining for Tiam-1 protein was observed in paracarcinoma gastric mucosa, obtained from patients with gastric cancer. By contrast, another study reported that positive staining for Tiam-1 protein was detected in gastric cancer tissues (11). In addition, as the degree of histological differentiation of gastric cancer decreased, the depth of invasion was found to increase, TNM stage was upgraded and lymph node metastasis was noted. Staining corresponding to Tiam-1 protein levels in gastric cancer tissues was observed to increase, indicating that Tiam-1 may represent a candidate biomarker of the invasive and metastatic state of gastric cancer and for patient prognosis; however, the intrinsic biological mechanism via which Tiam-1 exerts it effects remains unclear.

The aim of the present study was to further define the mechanism of gastric cancer cell invasion induced by Tiam-1. Firstly, two subclones with high (MH) or low (ML) invasive and metastatic potentials were isolated from the MKN-45 human gastric cancer cell line (M0) by in vitro adhesion selection, as described previously (12). Tiam-1 expression levels and cellular morphology were compared between the subclones. In addition, the correlation between Tiam-1 expression and the in vitro invasive potential of the three cell lines was analyzed. Next, the effects of Tiam-1 downregulation on changes in cytoskeletal structure and the invasiveness of MH cells were investigated by in vitro transfection of antisense oligodeoxynucleotides (ASODNs). Results of the present study are likely to provide an improved insight into the mechanism of gastric cancer cell invasion and metastasis induced by Tiam-1 and promote the development of novel treatment strategies in gastric cancer.

Materials and methods

Cell culture and adhesion selection

The parental cell line MKN-45 (M0) is a poorly differentiated gastric adenocarcinoma cell line (13) and was acquired from Shanghai Institute of Digestive Disease (Shanghai Second Medical University Renji Hospital, Shanghai, China). Frozen cell stocks were thawed and grown in RPMI-1640 medium (Hyclone Laboratories Inc., Logan, UT, USA) supplemented with 10% fetal calf serum (Gibco-BRL, Carlsbad, CA, USA), L-glutamine, sodium bicarbonate, essential amino acids, antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin) and sodium pyruvate. The medium was replaced with fresh medium every 2–3 days to maintain optimal growth at 37ºC in a humidified atmosphere containing 5% CO2. At ~80% confluence, cells were passaged at a ratio of 1:3 and detached using a 0.25% (w/v) trypsin-EDTA solution every 5–6 days.

M0 cells were selected for adhesion on mouse laminin-1-coated substrates (Sigma-Aldrich, St Louis, MO, USA), as described previously (12). Briefly, culture dishes (diameter, 35 mm) were coated with 50 μg laminin-1 in phosphate buffered saline (PBS) and air-dried overnight at 4ºC. Subconfluent M0 cells were incubated in 0.25% trypsin-EDTA solution, washed and resuspended in serum-free medium. The cell suspension (2.5×105 cells/dish) was incubated with laminin-1 coated substrate for 1 h and attached and unattached cells were collected separately. The attached cells were cultured to subconfluence on laminin-1 coated dishes, then selected again for adhesion. Unaffected cells were prepared following the same protocol. Serial adhesion selection was performed 20 times, then attached (MH) or unattached (ML) cells were maintained under normal culture conditions. Two passages after the last adhesion selection, the selected cells (MH and ML) were used for further experiments.

Antisense treatment of MH cell subclones

The design of one 18-mer ASODN was based on the nucleotide sequence of human Tiam-1, targeted to the DH domain, while sense oligonucleotide (SODN) was used as a control. Specific primers against ASODN and SODN were derived from previously published sequences (14) and determined using a BLAST search of a current EMBL database/NCBI-Genebank database (http://blast.ncbi.nlm.nih.gov/). The sequences used were as follows: SODN, 5′-*G*AT CTG CGA GCT CCT GG*A*-3′; and ASODN, 5′-*T*CC AGG AGC TCG CAG AT*C*-3′ (*represents phosphorothioate linkages). All ODNs were synthesized using the Applied Biosystems 391 DNA synthesizer (Perkin Elmer Applied Biosystems, Inc., Foster City, CA, USA).

MH cells in log-phase growth (2.5×105 cells) were plated onto culture dishes (diameter, 35 mm) and incubated under standard culture conditions for 24 h, then transfected with an ASODN or SODN of Tiam-1 using DOTAP liposomal transfection reagent (Roche GmbH, Mannheim, Germany). The preparation of liposome-ODN was performing according to the manufacturer’s instructions. Prior to treatment, ODNs and liposomes (ratio w/v, 5/15 μl; 1:3) were dissolved separately in HEPES-buffered saline solution (20 mm HEPES, 150 mm NaCl; pH 7.4) and mixed at room temperature for 15 min. The optimal working conditions for lipofectin-ODN were determined previously in preliminary experiments (data not shown). Experimental cells were treated with liposome-ASODN at a final concentration of 0.43 μM. Control treatment groups were serum-free culture medium only, liposome only (15 μl/dish) and liposome-SODN (0.43 μM). The culture medium was removed and cells were exposed to lipofection solution for 8 h. At the end of the incubation with ODN-liposome, the medium was replaced with fresh serum-containing culture medium (3 ml/dish) and placed into an incubator. Analysis of Tiam-1 mRNA expression in MH cells was performed 48 h following transfection.

Reverse transcription polymerase chain reaction (RT-PCR)

Total RNA was extracted using TriPure Isolation Reagent (Boehringer Mannheim, Mannheim, Germany) according to the manufacturer’s instructions, and its content and purity were evaluated by ultra-violet spectrophotography. cDNA was synthesized in 20-μl reaction volumes containing 1 μg total RNA, 0.2 μg/μl random hexamer primer, 4 μl 5X reaction buffer, 20 U/μl ribonuclease inhibitor, 10 mmol/l deoxyribonucleotide triphosphate (dNTPs) and 20 units Moloney murine leukemia virus reverse transcriptase (G-Biosciences & Geno Technology Co., St. Louis, USA). Procedures were performed according to the manufacturer’s instructions of the First Strand cDNA synthesis kit (Sangon Biological Engineering Co., Shanghai, China).

For the semi-quantitative assay, a standard calibration curve was created to determine the optimum number of cycles (15–35 cycles); the most appropriate number of cycles was 30 (data not shown). PCR was performed using the following materials: 2 μl cDNA, 5 μl 10X buffer, 4 μl MgCl2 (25 mmol/l), 8 μl dNTPs (2.5 mmol/l), 0.5 μl of each forward and reverse amplification primer (25 μmol/l) and 0.5 μl Ex Taq DNA polymerase (2.5 units; Takara Biotechnology Co., Ltd., Dalian, China) in a 50 μl reaction volume. Reactions were performed using the PTC-100TM PCR system (MJ Research Inc., Watertown, MA, USA) under the following conditions: 30 cycles of 94ºC (45 sec), 54ºC (1 min) and 72ºC (1.5 min). The following primer pairs (synthesized by Sangon Biological Engineering Co.) were used to amplify fragments of Tiam-1 (818 bp) and GADPH (310 bp) cDNAs: Tiam-1, forward 5′-TTC TCA CCA GTC TGT TCA GC-3′ and reverse 5′-CCA GAC TTG GAA TCC TCA GA-3′; and GADPH, forward 5′-AGG TCC ACC ACT GAC ACG TT-3′ and reverse 5′-GCC TCA AGA TCA TCA GCA AT-3′.

PCR products (5 μl) amplified from the Tiam-1 or GADPH genes were mixed and underwent 2% ethidium bromide agarose gel electrophoresis with molecular weight standards. The band intensities for Tiam-1 and GADPH were estimated using a Bio-Rad gel documentation 2000 system (Bio-Rad, Hercules, CA, USA). The intensity of Tiam-1 gene expression were reported as the ratio (RV) of Tiam-1 to GADPH (as an internal control). Specimens were analyzed three times and values were averaged.

Quantitative cell-ELISA

MH cells (1×105 cells) in log-phase growth were plated onto 96-well plates, incubated under standard culture conditions for 24 h and transfected as described. Control groups were prepared following the same protocol. Quantitative cellular-ELISA of the Tiam-1 protein in gastric cancer cells was performed following transfection for 48 h, as described previously (15).

In brief, culture medium was removed from the 96-well plates and the cells were washed twice with PBS, fixed with 4% paraformaldehyde (w/v in PBS) for 1 h and removed. Following three PBS washes, cells were blocked with 10% (w/v) goat serum albumin in PBS at 4ºC for 15 min and then incubated at 37ºC for 2 h with a 1:100 working dilution of rabbit polyclonal antibody (50 μl/well) against human Tiam-l (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Next, cells were washed 3 times with 0.05% Tween-20/PBS and incubated at 37ºC for 1 h with a 1:5,000 (50 μl/well) working dilution of horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Zhongshan Biotechnology Co., Ltd., Beijing, China). Following four washes with 0.05% Tween-20/PBS, equal amounts of A and B fluid (50 μl each) containing 3,3′,5,5′-tetramethylbenzidine peroxidase chromogenic reagent (KPL, Inc., Gaithersburg, MD, USA) were added to react in the dark for 10 min. The reaction was terminated with 50 μl/well HCl (1 mol/l) under oscillation for 30 sec and the absorbance at 450 nm was measured using an ELISA plate reader (Sunrise; Tecan Group, Ltd., Männedorf, Switzerland). Wells were washed with distilled water, followed by the addition of 0.08% (w/v) crystal violet (100 μl/well) and incubation at room temperature for 20 min. The wells were then washed again with distilled water and the reaction was terminated using 33% (w/v) glacial acetic acid (100 μl/well) at room temperature for 30 min and the absorbance was measured at 550 nm. The intensity of Tiam-1 protein was reported as the ratio (RD) of D450 to D550. Specimens were analyzed three times and the values were averaged. Negative controls consisted of omission of the primary antibody.

Morphological and ultrastructural observations

To examine morphological changes, cells were directly cultured on slides and washed twice with PBS. Next, the cells were stained with hematoxylin and eosin (H&E) and observed under a light microscope, and representative images were captured.

For scanning electron microscopy (SEM), cells were grown on cover slips and fixed with 2.5% (w/v) glutaraldehyde at 4ºC overnight. Following this, cells were washed with PBS, treated with 1% (w/v) osmium tetroxide (TAAB Laboratories Equipment Ltd., Berkshire, UK) at 4ºC for 1 h, washed three times with PBS and dehydrated using a gradient ethano1. SEM specimens were dried using a critical point dryer (Hcp-2; Hitachi Ltd., Tokyo, Japan) and the sputter was coated with Pt + Cd using an ion sputtering device (Oie-102; Hitachi Ltd.). Cell surface alteration was observed under an SEM microscope (AMRay 1000B, Bedford, MA, USA).

Cytoskeletal staining

Cytoskeletal staining was performed as described previously (16). Following two washes with PBS, cells cultured on slides were permeabilized with 0.1% Triton-X-100 3 times for 8 min, washed in M-buffered solution (50 mM imidazole, 50 mM potassium chloride, 0.5 mM magnesium chloride, 2 mM EGTA, 0.1 mM EDTA, 1 mM mercaptoethanol, 4 M glycerol and distilled water to 1,000 ml) 3 times for 4 min and then fixed with 3% (w/v) glutaraldehyde at room temperature for 10 min. Next, cells were washed in M-buffered solution 3 times for 3 min and stained with coomasssie brilliant blue solution (0.2 g R-250, 46.5 ml methanol, 7 ml glacial acetic acid and 46.5 ml distilled water) at room temperature for 40 min, followed by PBS washing, air-drying, dimethyl benzene clearing and neutral gum sealing. Cells were then viewed under a light microscope (LH50A; Olympus Optical Co., Tokyo, Japan).

In vitro invasion assay

In vitro cell invasion was analyzed using 24-transwell units (Millipore, Billerica, MA, USA), as described previously (17). Polyvinylpyrrolidone-free polycarbonate filters (diameter, 12 μm; pore size, 8 μm; Millipore) were coated with 30 μg Matrigel/filter (BD Biosciences, Franklin Lakes, NJ, USA). NIH-3T3 cell-conditioned media was used as a chemoattractant and was obtained by culturing cells for 24 h in serum-free RPMI-1640 medium. Cells in serum-free RPMI-1640 medium (400 μl containing 1×105 cells) were added to the upper compartment of the chamber, while 600 μl conditioned medium was added to the lower compartment. Following 72 h incubation at 37ºC in humidified 95% air with 5% CO2, basement membrane Matrigel and cells on the upper side of the filter were removed by wiping with a cotton swab. Cells that had migrated to the lower surface of the filters were fixed and stained with H&E. The invasive potential of cells was determined by measuring the number of cells that migrated to the lower side of the filters through the basement membrane matrigel and pores. The number of cells that penetrated the filter was counted in 10 microscopic fields of each filter. Assays were performed in triplicate.

Statistical analysis

Quantitative data are presented as the mean ± SD. All data were analyzed statistically by one-way ANOVA followed by a least significant difference test, using the SPSS 10.0 software program (SPSS Inc., Chicago, IL, USA). P<0.05 was considered to indicate a statistically significant difference. Correlations were calculated using the Spearman correlation coefficient.

Results

In vitro invasive potential of M0, ML and MH cells

MH and ML cells, which exhibit different invasive and metastatic potentials, were subcloned from the MKN-45 human gastric cancer cell line (M0) by adhesion selection. A transwell assay was performed to compare the invasive capacity of the three cell strains in vitro. As demonstrated in Table I and Fig. 1, MH cells exhibited a higher invasive potential than M0 and ML cells in vitro. This is consistent with results of a previous study on tumor-cell transplantation and experimental metastasis in nude mice (12).

Figure 1 In vitro invasive potential of M0, ML and MH cells. *P<0.05 vs. M0; ΔP<0.05 vs. ML. MO, MKN-45 human gastric cancer cell line; MH, cell subpopulation with a high invasive and metastatic potential; ML, cell subpopulation with a low invasive and metastatic potential.

Table I In vitro invasive potential of gastric cancer cells.

Table I

In vitro invasive potential of gastric cancer cells.

Variable	Cells (n)			F-value
	M0	ML	MH
Cells/field	11.67±3.79	9.67±3.06	24.33±8.02a,b	6.470

{ label (or @symbol) needed for fn[@id='tfn1-mmr-08-02-0425'] } Data are presented as the mean ± SD.

a P<0.05 vs. M0;

b P<0.05 vs. ML.

{ label (or @symbol) needed for fn[@id='tfn4-mmr-08-02-0425'] } MO, MKN-45 human gastric cancer cell line; MH, cell subpopulation with a high invasive and metastatic potential; ML, cell subpopulation with a low invasive and metastatic potential.

Tiam-1 mRNA and protein expression levels in M0, ML and MH cells

Since Tiam-l has been previously reported to be associated with tumor invasion and metastasis, Tiam-l mRNA and protein expression levels in M0, ML and MH cells were analyzed by RT-PCR and quantitative cellular-ELISA. The expression of Tiam-1 mRNA and protein in the three strains was positive; however, levels in MH cells were observed to be markedly higher than those of M0 and ML cells (Table II; Figs. 2–4)

Figure 2 RT-PCR analysis demonstrated that the MH cell subclone expressed Tiam-1 as a marked 818 bp-band, while M0 and ML cells expressed faint bands. RT-PCR, reverse transcription polymerase chain reaction; MO, MKN-45 human gastric cancer cell line; MH, cell subpopulation with a high invasive and metastatic potential; ML, cell subpopulation with a low invasive and metastatic potential; Tiam-1, T-cell lymphoma invasion and metastasis-inducing factor 1.

Figure 4 Tiam-1 protein expression levels among M0, ML and MH cells. *P<0.05 vs. M0; ΔP<0.05 vs. ML. Tiam-1, T-cell lymphoma invasion and metastasis-inducing factor 1; MO, MKN-45 human gastric cancer cell line; MH, cell subpopulation with a high invasive and metastatic potential; ML, cell subpopulation with a low invasive and metastatic potential; RD, ratio of D450 to D550.

Table II Tiam-1 mRNA and protein expression levels in gastric cancer cells.

Table II

Tiam-1 mRNA and protein expression levels in gastric cancer cells.

Ratio	Cells			F-value
	M0	ML	MH
RV	0.759±0.047	0.743±0.039	0.855±0.051a,b	5.269
RD	0.911±0.104	0.892±0.101	1.262±0.165a,b	13.429

{ label (or @symbol) needed for fn[@id='tfn5-mmr-08-02-0425'] } Data are presented as the mean ± SD.

a P<0.05 vs. M0;

b P<0.05 vs. ML.

{ label (or @symbol) needed for fn[@id='tfn8-mmr-08-02-0425'] } Tiam-1, T-cell lymphoma invasion and metastasis-inducing factor 1; R_D, ratio of D450 to D550; R_V, ratio of Tiam-1 to GADPH; MO, MKN-45 human gastric cancer cell line; MH, cell subpopulation with a high invasive and metastatic potential; ML, cell subpopulation with a low invasive and metastatic potential.

These observations were analyzed further, revealing a positive correlation between the expression levels of Tiam-l mRNA or protein and the invasive capacity of the cells (Spearman correlation coefficient, r=1.000; P<0.01).

Antisense treatment downregulates the expression of Tiam-1 mRNA and protein in MH cells

To confirm the effect of Tiam-1 ASODN on MH cells, RT-PCR and quantitative cell-ELISA were performed. The optimum conditions for antisense treatment, determined in preliminary experiments (data not shown), were 0.43 μM ODN, ratio (w/v) of ODN/liposome=1:3, 4 h exposure and 48 h recovery. Following ASODN-liposome treatment, the expression of Tiam-1 in MH cells was significantly inhibited compared with untreated (control), liposome- or SODN-liposome-treated samples (Table III). RT-PCR revealed a faint Tiam-1 band in MH cells following ASODN treatment. By contrast, the band was significantly stronger in the other three groups. No significant difference between the intensities of the stronger bands was identified (Figs. 5 and 6). Quantitative cell-ELISA of MH cells transfected with ASODN-liposome also revealed a significant reduction in Tiam-1 protein levels compared with untreated, liposome- or SODN-liposome treated groups (Fig. 7).

Figure 5 RT-PCR analysis revealed that ASODN-liposome (AS-L)-treated MH cells faintly expressed Tiam-1 as an 818-bp band. Untreated (control, C), liposome only-treated (L) and SODN-liposome (S-L)-treated MH cells expressed strong bands. RT-PCR, reverse transcription polymerase chain reaction; SODN, sense oligonucleotide; ASODN, antisense oligodeoxynucleotide; Tiam-1, T-cell lymphoma invasion and metastasis-inducing factor 1; MH, cell subpopulation with a high invasive and metastatic potential.

Figure 6 Tiam-1 mRNA expression levels following antisense treatment. *P<0.05 vs. ASODN-liposome-treated MH cells. Tiam-1, T-cell lymphoma invasion and metastasis-inducing factor 1; SODN, sense oligonucleotide; ASODN, antisense oligodeoxynucleotide; MH, cell subpopulation with a high invasive and metastatic potential; RV, ratio of Tiam-1 to GADPH.

Figure 7 Tiam-1 protein expression levels following antisense treatment. P<0.05, vs. ASODN-liposome treated MH cells. Tiam-1, T-cell lymphoma invasion and metastasis-inducing factor 1; SODN, sense oligonucleotide; ASODN, antisense oligodeoxynucleotide; MH, cell subpopulation with a high invasive and metastatic potential; RD, ratio of D450 to D550.

Table III Tiam-1 mRNA and protein expression levels in gastric cancer cells.

Table III

Tiam-1 mRNA and protein expression levels in gastric cancer cells.

Ratio	Control	Liposome	SODN-liposome	ASODN-liposome	F-value
RV	0.801±0.065a	0.789±0.054a	0.754±0.039a	0.162±0.018	130.160
RD	1.290±0.182a	1.237±0.108a	1.234±0.103a	0.982±0.119	5.516

{ label (or @symbol) needed for fn[@id='tfn9-mmr-08-02-0425'] } Data are presented as the mean ± SD.

a P<0.05 vs. ASODN-liposome-treated MH cells.

{ label (or @symbol) needed for fn[@id='tfn11-mmr-08-02-0425'] } Tiam-1, T-cell lymphoma invasion and metastasis-inducing factor 1; R_D, ratio of D450 to D550; R_V, ratio of Tiam-1 to GADPH; SODN, sense oligonucleotide; ASODN, antisense oligodeoxynucleotide; MH, cell subpopulation with a high invasive and metastatic potential.

Alterations in cellular morphology and ultrastructure following antisense treatment

Following H&E staining, cells were observed under a light microscope. M0 and ML cells were floating and exhibited round epithelial-like morphologies with a large nucleocytoplasmic ratio; no differences in morphology were found between M0 and MH cells. However, MH cells were observed to be multiform compared with the uniform appearance of M0 and ML cells. A number of MH cells were observed to exhibit a polygonal morphology with longer cytoplasmic processes (Fig. 8). Following Tiam-1 ASODN treatment, the morphology of MH cells resembled that of M0 and ML cells.

Figure 8 Gastric cancer cell morphology. Gastric cancer cells stained with H&E were observed by light microscopy. (A) M0, (B) ML, (C) MH and (D) antisense-treated MH cells (magnification, ×200). H&E, hematoxylin and eosin; MO, MKN-45 human gastric cancer cell line; MH, cell subpopulation with a high invasive and metastatic potential; ML, cell subpopulation with a low invasive and metastatic potential.

SEM revealed abundant microvilli, cytoplasmic processes and large pseudopodia dotted on the surface of MH cells. Certain processes extended a considerable distance, radiated in all directions or formed bulges at their ends. By contrast, the processes of M0 and ML cells were short and compact. No desmosomes, tight conjunctions or other cell junction structures were observed in the three cell types. Following antisense treatment, MH cells exhibited a smooth surface with markedly reduced filopodia and microspikes (Figs. 9 and 10).

Figure 9 Scanning electron microscopy. Ultrastructural features of gastric cancer cells. MO, MKN-45 human gastric cancer cell line; MH, cell subpopulation with a high invasive and metastatic potential; ML, cell subpopulation with a low invasive and metastatic potential. Magnification, ×1,200.

Figure 10 Scanning electron microscopy. Ultrastructural alterations of MH cells following antisense treatment. MH, cell subpopulation with a high invasive and metastatic potential; AS-MH, antisense treatment MH. Magnification, ×6,000.

Cytoskeletal structure and reorganization following antisense treatment

Cells were stained with coomasssie brilliant blue R-250 and viewed under an immersion objective. The cytoskeletal structure of M0, ML and MH cells formed a complicated cytoskeletal network in the cytoplasm and was markedly more concentrated in perinuclear areas. Bundles of cytoskeletal filaments were stretched and scattered intercellularly. Compared with M0 and ML cells, cytoskeletal alignment in MH cells was disordered, accompanied with more dotted actin bodies, cytoplasmic processes and long filament-like structures. By contrast, no differences in cytoskeletal alignment were observed between M0 and ML cells. Following antisense treatment, cytoskeletal distribution in MH cells changed from irregular to regular, with reduced long filament-like structures on the cell surface and dotted actin bodies in the cytoplasm, and resembled that of M0 and ML cells (Fig. 11).

Figure 11 Coomasssie brilliant blue staining of cytoskeletal changes in (A) M0, (B) ML, (C) MH and (D) ASODN-treated MH cells. Images captured under an immersion objective (magnification, ). MO, MKN-45 human gastric cancer cell line; MH, cell subpopulation with a high invasive and metastatic potential; ML, cell subpopulation with a low invasive and metastatic potential; ASODN, antisense oligodeoxynucleotide. Magnification, ×1,000.

Tiam-1 ASODN treatment downregulates the invasiveness of MH cells in vitro

An in vitro transwel1 assay revealed that the number of antisense-treated MH cells penetrating the artificial basement membrane was 10.33±2.52 cells/high power field, which was significantly lower than that of untreated MH cells (25.67±6.11 cells/high power field; P<0.05; Table IV, Fig. 12).

Figure 12 In vitro invasive potential of ASODN-treated or untreated MH cells. *P<0.05 vs. ASODN-treated MH cells. ASODN, antisense oligodeoxynucleotide; MH, cell subpopulation with a high invasive and metastatic potential.

Table IV Effect of antisense treatment on the in vitro invasive potential of MH cells.

Table IV

Effect of antisense treatment on the in vitro invasive potential of MH cells.

Variable	Cells (n)		F-value
	Untreated	ASODN-treated
Cells/field	25.67±6.11	10.33±2.52a	16.153

{ label (or @symbol) needed for fn[@id='tfn12-mmr-08-02-0425'] } Data are presented as the mean ± SD.

a P<0.05 vs. ASODN-treated MH cells.

{ label (or @symbol) needed for fn[@id='tfn14-mmr-08-02-0425'] } ASODN, antisense oligodeoxynucleotide; MH, cell subpopulation with a high invasive and metastatic potential.

Discussion

Tumor cell invasion and metastasis depends on a unique set of biological properties that enable malignant cells to complete all steps of the metastatic cascade. A tumor mass commonly consists of various cell subclones, a phenomenon known as tumor heterogeneity (18). Since subclones are from the same parental cell line, thus having the same genetic background, the differences in their phenotypes are likely to have a number of underlying molecular mechanisms (19,20). In-depth study of the differences between subclones may provide novel insight into the mechanisms of tumor cell invasion and metastasis.

Members of the Rho GTPase family have emerged as key players in the regulation of a diverse set of biological activities, including actin organization, focal complex/adhesion assembly, cell motility, cell polarity and gene transcription, the importance of which is becoming evident in cancer progression, particularly in the area of metastasis (21,22). As an upstream regulator of Rho-GTPases, particularly Rac-1, the potential role of Tiam-1 in tumor invasion and metastasis is becoming increasingly apparent.

In the present study, two subclones with a high (MH) or low (ML) invasive and metastatic potential were acquired from the MKN-45 human gastric cancer cell line (M0) using in vitro adhesion selection on laminin-1-coated dishes (12), and Tiam-1 expression levels were compared. Results indicated that Tiam-1 mRNA and protein expression levels in MH cells were markedly higher than that of M0 and ML cells. In addition, a positive correlation between the expression of Tiam-1 and the in vitro invasive potentials of M0, ML and MH cells was observed. These results are consistent with previous observations (11) and implicate a role for Tiam-1 in the invasive process of human gastric cancer.

Cytoskeletal components, including microtubules, microfilaments and intermediate filaments, are responsible for the maintenance of cell morphology, direction of cell locomotion, cell spreading, cell migration and stability of cell-substrate contact in normal cells (23). Alterations in cytoskeletal structures, such as reduced microtubules, disruption of stress fibers and redistribution of actin-filaments in tumor cells, have been reported to be important determinants of malignancy in tumor cells, enabling cells to transform and transmigrate within tissues (24). In the present study, analysis of the cellular morphology and ultrastructure of gastric cancer cells revealed that M0 and ML cells were floating and exhibited round epithelial-like structures. By contrast, MH cells were multiform and a number of cells were polygonal with long cytoplasmic processes. SEM also revealed abundant microvilli, cytoplasmic processes and large microhills located on the MH cell surface. Specific processes were observed to extend considerable distances, radiate in all directions or form bulges at their ends. By contrast, M0 and ML cells were observed to exhibit short and compact processes. Following staining with coomassie brilliant blue, the cytoskeletal structure of M0, ML and MH cells was observed to form a complicated cytoskeletal network in the cytoplasm. The network was particularly concentrated in perinuclear areas and bundles of cytoskeletal filaments passed through the cell wall to interact with cytoskeletal arrays of adjacent cells. Compared with M0 and ML cells, cytoskeletal alignment in MH cells was disordered, accompanied with increased dotted actin bodies, cytoplasmic processes and long filament-like structures, whereas no marked differences were observed between M0 and ML cells. The differences in cytoskeletal structure may be due to cytodynamic activities associated with dissemination and implantation, based on the invasive and metastatic potentials of M0, ML and MH cells. As Tiam-1 is known to regulate the reorganization of cytoskeletal structure, differences in the morphological characteristics of gastric cancer cells with different invasive and metastatic potentials may correlate with Tiam-1 expression.

ASODNs bind and inactivate specific RNA sequences and represent one of the best tools for studying gene function, the regulation of gene expression and interactions between gene products. Intracellular delivery of ASODNs results in the downregulation of target gene expression by inhibiting transcription or translation without affecting other cellular functions. To validate the role of Tiam-1 in gastric cancer cell invasion, a 18-mer ASODN was used to block Tiam-1 expression in MH cells and the subsequent changes in cellular morphology and the in vitro invasiveness of MH cells were analyzed. As unmodified phosphodiester oligomers are susceptible to nucleases, partial phosphorothioate oligomers were used, which are relatively nuclease-resistant. Results demonstrated that the treatment of MH cells with ASODNs (0.43 μM, 48 h) decreased Tiam-1 mRNA transcription and protein expression by 70 [(0.801–0.162)/0.801] and 65% [(1.290–0.982)/1.290] respectively, compared with untreated controls. In addition, the in vitro invasive potential of ASODN-treated MH cells was suppressed by 75% [(25.67–10.33)/25.67]; however, cells treated with SODN or lipofectin alone were not affected. Morphological and ultrastructural observations also revealed that ASODN-treated MH cells exhibited a smooth surface with markedly reduced filopodia and microspikes, which resembled M0 and ML cells. Cytoskeletal distribution in ASODN-treated MH cells was markedly altered from disordered to regular, with reduced long filament-like structures, projections, pseudopodia at the cell surface and reduced actin bodies in the cytoplasm. Our findings showed that Tiam-1 induces gastric cancer cell invasion by regulating cytoskeletal reorganization, which caused the depolymerization of microtubules and microfilaments, leading to disordered, actively motile properties, cell detachment and a metastatic potential in gastric cancer cells. Previous observations showed that Tiam-1 contributed to the cytoskeletal reorganization required during cell migration and neurite extension by activation of Rac 1 (25). Whether this is the case in gastric cancer cell invasion it is not yet known.

In addition, abnormalities in the expression and functional activity of cell adhesion molecules are implicated in the invasive and metastatic progression of tumor cells. Intercellular and cell-matrix adhesion molecules have been shown to have roles in addition to acting as cementing substances, such as regulating cell polarity, differentiation, invasion and migration (26). A number of these cellular events are mediated through direct associations with the cytoskeletal network and interactions between the cytoskeletal network. Therefore, the manipulation of molecules associated with the cytoskeleton requires further investigation (27). The in vitro cell model of laminin-1 adhesion selection used in the present study represents a simple and time-effective method to screen a specific phenotypic cell subpopulation and is likely to be useful for future analysis of the cytoskeleton.

Laminin-1, a major basement membrane glycoprotein, promotes the malignant phenotype and expression of specific laminin receptors, and has been found to correlate with the malignant characteristics of tumor cells (28). The adhesion of cells to laminin-1 induces collagenase IV production, which is required for tumor cells to metastasize. Previous studies, as well as the present study, have demonstrated that the adherent subclone (MH), isolated from the MKN-45 gastric cancer cell line by in vitro adhesion selection on laminin-1, is more invasive and metastatic in vitro and in vivo compared with parental (M0) and non-adherent subclone (ML) cells, indicating that the properties of these cells are closely associated with adhesion to the basement membrane and extracellular matrix (12). In addition, other studies have demonstrated that a laminin-adhesive subclone of a human colon cancer cell line revealed higher invasive and metastatic potentials in vitro and in vivo than a laminin-non-adhesive cell line. These observations indicate that the increased invasive and metastatic potentials of the laminin adhesion-selected subpopulation may be due to an alteration in the membrane distribution and/or affinities of multiple laminin receptors, including the 67 kDa laminin receptor and β1 integrin (29,30). The cancer cell population, either as a solid tumor mass in vivo or as a continuous cell line in vitro, is an ever-changing entity due to their genetic instability and selective environmental pressure (31). Tiam-1 expression is associated with the invasive potential of gastric cancer cells. Therefore, we hypothesized that Tiam-1 expression, the rearrangement of cytoskeletal structure and the expression and/or redistribution of cell surface adhesion molecules are linked. Additionally, the associated mechanisms of cellular signal transduction remain unclear and require investigation.

Results of the present study demonstrated that Tiam-1 induces gastric cancer cell invasion and may represent a candidate target for biotherapy. However, additional studies are required to determine the molecular and biological mechanisms associated with these observations.

Acknowledgements

The authors thank Professor Zheng Jiang and his personnel at the Molecular Biological Center of Xinan Hospital (Chongqing, China) for their excellent technical assistance.

References