All human esophageal squamous carcinoma cell lines KYSE 170 parental (wild) and derivate cells (KYSE 170mock, KYSE 170F3 and KYSE 170F4) were established in our laboratory and maintained in RPMI 1640 (Life Technologies, Gaithersburg, MD) and Ham's F12 (Nissui Pharmaceutical, Tokyo, Japan) with 2% fetal bovine serum (FBS).

Whole-cell extract lysate was prepared from 1 × 10⁷ cells in a sample buffer (2% sodium dodecyl sulfate [SDS], 10% glycerol, 50 mM Tris-Hcl, pH 6.8) at room temperature. Cell lysates were sonicated and the protein content was measured with the Bradford method using BCA Protein Assay Reagent (Pierce, Rockford, MA), after which the cell lysates were electrophoresed on a 12% polyacrylamide gel SDS page and transferred to a polyvinylidene difluoride membrane (Immobilon, Milipore, Bedford, MA) using a semidry transfer blot system (Bio-Rad, Hercules, CA). The membranes were blocked with TBS (20 mM Tris, 150 mM NaCl, pH 7.6) containing 1% Tween 20 and 5% skimmed milk for 1 hour. The membranes were incubated overnight at 4°C with anti-human fascin mouse monoclonal antibody (DAKO, Osaka, Japan; diluted 1:500) or with anti-human β-actin mouse monoclonal antibody (Sigma Inc., St. Louis, MO; diluted 1:2000) as an internal control. They were washed and then incubated at room temperature for 1 hour with anti-mouse IgG-HRP (Zymed, San Francisco, CA), as a secondary antibody, and analyzed using ECL plus reagent (Amersham, Buckinghamshire, UK).

Total RNA was extracted from KYSE cell lines by the TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocols 2829. Reverse transcription of total cellular RNA was performed using a First-Strand cDNA Synthesis Kit (Amersham, Buckinghamshire, UK). cDNA was subjected to PCR for 25 cycles of amplification using an Advantage cDNA PCR kit (Becton Dickinson Biosciences, Palo Alto, CA). Each PCR cycle consisted of a denaturation step for 1 minute at 94°C and an annealing step for 1 minute at 60°C. The final extension step was carried out for 5 minutes at 72°C. The PCR products were separated on 1.5% agarose gel and visualized by ethidium bromide staining. PCR primers used for fascin were 5'- AGGCGGCCAACGAGAGGAAC-3' as the forward primer and 5'-ACGATGATGGGGCGGTTGAT-3' as the reverse primer; and for glyceraldehydes-3-phosphate dehydrogenase (G3PDH), 5'-TGGTATCGTGGAAGGACTCATGAC-3' was used as the forward primer and 5'-ATGCCAGTGAGCTTCCCGTTCAGC-3' as the corresponding reverse primer. cDNA from HeLa cells was used as a positive control for each analysis.

Cells were plated into 6 cm dishes (2 × 10⁴ cells per dish) at day 0 and incubated for 24 hours for sufficient cell growth. Cells were harvested with trypsin/EDTA every 48 hours for five days, and counted with a cell counter (Coulter Z1, Beckman Coulter, Fullerton, CA). To examine the effect of the suppression of fascin expression on cell growth, we compared it with the control culture in triplicate. Each experiment was repeated three times independently.

Cells were seeded into collagen-coated plates (Becton Dickinson, MA) without FBS. After 24 hours of incubation, the adherent and floating cells were counted from five randomly selected fields. The assay was repeated three times under the same conditions.

200 μmol/L Z-VAD-FMK (BD Biosciences, San Jose, CA) was added to the cells after the cells were plated. The inhibition of cell growth was measured by MTT assay and cell counting assay. The assay was repeated three times under the same conditions.

In order to construct a vector for fascin-small interfering RNA (siRNA), pSilencer2.1-U6 hygro (Ambion, Inc., Austin, TX) was digested with BglII and HindIII. A chemically synthesized oligonucleotide encoding a fascin-short hairpin siRNA that included a loop motif was inserted downstream of the U6 promoter of the plasmid using a DNA ligation kit (Takara Bio, Inc., Shiga, Japan), and was cloned. Sequences of the oligonucleotide targeted to fascin were 5'-GCCUGAAGAAGAAGCAGAU-3' corresponding to positions 116 to 134 within fascin exon 1. An ESCC cell line KYSE170 was stably transfected with either the fascin-siRNA expression vector or the negative control vector (pSilencer2.1-U6 hygro) using FuGene6 reagent (Roche Diagnostics, Basel, Switzerland) 26, and cell clones were selected against 100 μg/mL hygromycin (Nacalai Tesque, Kyoto, Japan).

Suspensions of 1 × 10⁶ KYSE 170 parental (wild) and derivate cells (KYSE 170mock, KYSE 170F3 and KYSE 170F4) in PBS (60 μL) were injected subcutaneously into the right flanks of 5-week-old male BALB/c nu/nu mice (Japan SLC, Shizuoka, Japan) at day 0. The inoculation was conducted in five mice per group, and tumor growth was estimated from the average volume of tumors by the formula: 1/2 × L² × W (where L is the length of the tumor and W is the width of the tumor). At 30 days after injection, all mice were sacrificed and the subcutaneous tumors were resected and fixed in 10% formaldehyde/PBS. Tumors were then embedded in paraffin and stained with H&E and fascin. All animal experiments were performed in accordance with institutional guidelines.

Resected tumors from the in vivo experiments were fixed in 10% formaldehyde solution, embedded in paraffin, cut in 4 μm thick sections, and mounted on aminopropyltriethoxylane-coated glass slides. Immunohistochemical staining was carried out using an Envision Kit (Dako Cytomation, Glostrup, Denmark). Tissue sections were incubated overnight at 4°C with anti-human fascin monoclonal antibody clone 55kDa (DAKO, Osaka, Japan; 1:50 dilution) and then incubated with biotinylated anti-mouse IgG for 30 minutes at room temperature. Tissue sections were then stained with 3,3' diaminobenzidine liquid system (Dako Cymation, Glostrup, Denmark), counterstained with Mayer's haematoxilyn, dehydrated and mounted. We performed the same protocol using goat anti-human caspase 3 polyclonal antibody (Santa Cruz Biotechnology, CA, USA; dilution 1:500). As a negative control, the primary antibody was replaced with normal mouse IgG, and a further control was carried out without the primary antibody.

The DeadEnd Colorimetric TUNEL system kit was used for the TUNEL assay. The paraffin-embedded tissues were fixed in 4% formaldehyde with PBS for 15 minutes, permeabilized with 20 μg/ml proteinase K for 15 minutes and subsequently incubated at 37°C for 60 minutes with the rTdT reaction mix on the slide. Streptavidin HRP solution 1:500 in PBS was added and incubated for 30 minutes. DAB solution was then added for 5 minutes. Then the apoptotic cells were counted in five different fields at 40 ×, obtaining the average and standard deviation (SD).

The statistical analysis was performed using the software StatView 4.5 (Abacus Concept, Berkley, CA). A p-value of < 0.05 indicated statistical significance.

The expression of fascin was reduced by 74.8% in the stable subclones 170F1, 89.1% in 170F2, 93% in 170F3 and 97.7% in 170F4. The transfection for the nonspecific siRNA vector (170 mock) did not affect the expression of fascin. HeLa cells were used as a positive control for fascin (Figure 1A). Semiquantitative reverse transcription-PCR was performed, obtaining similar results to western blot analysis (Figure 1B).

We hypothesized that knockdown of fascin might affect cell growth. In order to verify this hypothesis, we investigated the effect of fascin knockdown on proliferation in one ESCC cell line (KYSE 170). With the suppression of fascin, the cell growth was inhibited by 40.3% in KYSE 70F4 at the day 5, as compared to the empty vector siRNA transfected cell line (KYSE 170 mock) (p < 0.01) (Figure 2).

In the cell adhesion in vitro experiment, the differences between KYSE 170 mock and fascin knockdown cells were not observed until 6 hours after seeding. The KYSE 170 mock cells attached well after 6 hours, whereas the fascin knockdown cells (KYSE 170F4) did not attach properly and an increasing number of cells was floating. After 24 hours of seeding, the number of adherent living cells was significantly lower in the KYSE 170F4 group than in the KYSE 170 mock group (Figure 3A); in contrast, the number of floating cells increased (Figure 3B). As a result, the total number of cells decreased in the fascin knockdown group. These results suggest that fascin knockdown induces the loss of cell adhesion to the matrix and might result in the inhibition of cell growth in ESCC cells.

The general caspase inhibitor (Z-VAD-FMK) did not affect the inhibition of attachment induced by fascin knockdown (Figures 3C and 3D). The difference in cell count at 72 hours between the fascin knockdown cell line and the mock cell line remained the same, both using the general caspase inhibitor and without (mean 40.1% both). These results provide convincing evidence that detachment may be caused by fascin knockdown.

Using the in vitro model and western blot analysis, we observed a decreasing expression of the two proteins involved in attachment and cell cycle signaling pFAK and Integrin α2 in all the studied clones, with the highest expression in KYSE 170wild and mock, and the lowest expression in KYSE 170F4 cells (Figure 4). These results suggest that the occurrence of interactions between fascin and other proteins or membrane proteins may occur.

In order to investigate the possible effect of fascin-siRNA on tumor formation in vivo, we performed a subcutaneous tumor formation assay in nude mice (Figure 5A). We evaluated tumor volume and weight at 30 days after inoculation. The tumor formations caused by fascin-siRNA transfected cells (KYSE 170F3 and KYSE 170F4) were evidently smaller than those formed by KYSE 170wild and KYSE 170 mock (Figure 5B). The average tumor volume was significantly lower in KYSE 170F3 and KYSE 170F4 by 91% and 95%, respectively, compared to KYSE 170 mock (p < 0,0001 and p < 0,0001, respectively). The average tumor weight was also inhibited in KYSE 170F3 and KYSE 170F4 by 87% and 90%, respectively (p < 0,0004 and p < 0,0003, respectively). Finally, all tumors were stained with H&E and fascin, and fascin protein expression was reduced to a minimum in tumors of inoculated mice with fascin-siRNA transfected cells, although not in tumors of empty vector transfected cells (KYSE 170 mock) (Figure 6).

The analysis of tumoral apoptosis was conducted by TUNEL assay (Figure 7A) and immunohistochemistry for caspase 3 (Figure 7B) where the number of apoptotic cells was higher in the fascin downregulated tumor than in the KYSE 170 mock tumor. For the TUNEL assay, the average of apoptotic cells in the fascin knockdown tumor was 15 ± 3.6 cells per high power field and for the KYSE 170 mock tumor it was 2 ± 0.83 cells per high power field.