Involvement of TSLC1 in Progression of Esophageal Squamous Cell Carcinoma

INTRODUCTION

Esophageal carcinoma is the sixth frequent cause of cancer death in the world (1) , and ESCC 3 accounts for >90% of the esophageal carcinoma in Asian countries. Although surgical techniques and perioperative management have progressed, the prognosis for patients with ESCC remains poor. Finding molecular therapeutic targets for ESCC treatment is one of the most promising avenues of research that might help to improve the survival of patients with this type of refractory cancer. Some of the genetic alterations associated with development or progression of ESCC have been described (2) . However, few of these genes have been demonstrated to be associated with biological or pathological features of ESCC. Therefore, novel genes associated with a progression of ESCC apparently need to be identified.

Frequent allelic losses of 11q23 in ESCC were reported previously, but no tumor suppressor genes in this locus have been identified in ESCC. Comparative genomic hybridization analysis revealed that 11q23 was lost in 10 of 29 (34%) of ESCC cell lines (3) and 17 of 46 (37%) of ESCC tumor samples (4) . TSLC1 was identified on chromosome 11q23.2 as a tumor suppressor gene in NSCLC by functional complementation of a lung adenocarcinoma cell line, A549, through suppression of tumorigenicity in nude mice (5) . TSLC1 encodes a transmembrane glycoprotein of 442 amino acids. The protein has structural homology to the extracellular domains of neural cell adhesion molecules and is suggested to be involved in cell–cell adhesion (6 , 7) . TSLC1 expression is reduced or absent in several cancer cell lines, including NSCLC, hepatocellular carcinoma, and pancreatic cancer. Promoter hypermethylation is suggested as a major mechanism of silencing the TSLC1 gene in these cancers (5 , 8 , 9) . On the other hand, inactivating mutations of TSLC1 have been rarely reported (10 , 11) .

Some studies have demonstrated that TSLC1 is involved in the progression of several types of cancer, whereas the role of TSLC1 in ESCC has not yet been examined. Although the restoration of TSLC1 expression suppressed the tumorigenesis and metastasis of an NSCLC cell line (5 , 12) , the mechanisms underlying these phenotypic changes, including growth suppression, are still unknown. In addition, the clinical significance of TSLC1 expression was not demonstrated in the previous reports. In this study, to elucidate the possible involvement of TSLC1 and evaluate its clinical significance in ESCC, we examined the expression and methylation status of TSLC1 in ESCC and analyzed the clinicopathological as well as biological features of TSLC1 alteration in ESCC.

MATERIALS AND METHODS

Tumor Samples and Cell Culture.

Tumor samples and corresponding normal tissues were obtained from 56 patients with primary ESCC who underwent surgery at Kyoto University Hospital from 1990 to 1997; the observation period ranged from 2 to 140.4 months (the median period was 71 months). None of these patients received preoperative treatment, such as chemotherapy or radiotherapy. All of the tumors were confirmed as ESCC by the clinicopathological department of the hospital. All of the cases were classified according to the fifth edition of the pathological tumor node metastasis classification of 1997 (13) . Informed consent for the use of their samples was obtained from all of the patients before surgery. Thirty-six ESCC cell lines used in this study were established in our laboratory and maintained in RPMI 1640 (Invitrogen, Carlsbad, CA) and Ham’s F12 (Nissui Pharmaceutical, Tokyo, Japan) mixed (1:1) medium containing 2% fetal bovine serum (14) . Two normal esophageal epithelial cell lines (HEEC1 and KNEC2) were established in our laboratory and maintained in keratinocyte serum-free medium containing 2.5 μg of epidermal growth factor and 25 μg of bovine pituitary extract (Invitrogen; Refs. 15 and 16 ).

Expression Vectors and an Antibody.

The plasmids pcTSLC1 and pcDNA3.1-Hygro (+; Invitrogen) were used for the transfections in this study. The construction of pcTSLC1 was described previously (5) . Rabbit anti-TSLC1 polyclonal antibody CC2 was used for Western blotting and immunofluorescence staining as described previously (6 , 12) .

Purification of Total Cellular RNA and RT-PCR.

Total cellular RNA was purified from cell lines and frozen stored tissues of ESCC patients by the acid guanidine-phenol-chloroform method. Reverse transcription of total cellular RNA (5 μg) was performed using a First-Strand cDNA Synthesis Kit (Amersham, Buckinghamshire, United Kingdom), and cDNA was subjected to PCR for 35 cycles of amplification using Advantage cDNA PCR kit (Becton Dickinson Biosciences, Palo Alto, CA). Amplification was performed for 30 s at 94°C and 3 min at 68°C. The final extension step was carried out for 3 min at 68°C. The amplification products were separated on 1.5% agarose gels and visualized by ethidium bromide staining. PCR primers for TSLC1 were as follows: the forward primer was 5′-CATCACAGTCCTGGTCCCACCACGTAATCT-3′, and the reverse primer was 5′-AATAGGGCCAGTTGGACACCTCATTGAAAC-3′. For glyceraldehyde-3-phosphate dehydrogenase, the forward primer was 5′-TGGTATCGTGGAAGGACTCATGAC-3′, and the reverse primer was 5′-ATGCCAGTGAGCTTCCCGTTCAGC-3′. For the positive controls, normal human tissue cDNA of lung were purchased from Becton Dickinson Biosciences (Palo Alto, CA).

Northern Blot Analysis.

Poly(A) RNA was purified from total cellular RNA using the Oligotex-dT30(super; Takara Bio, Shiga, Japan). A 961-bp PCR-derived fragment was used as a probe for detection of TSLC1. PCR primers were as follows: forward primer was 5′-CATCACAGTCCTGGTCCCACCACGTAATCT-3′, and the reverse primer was 5′-AATAGGGCCAGTTGGACACCTCATTGAAAC-3′. Poly(A) RNA (0.5 μg) was electrophoresed on 1% agarose-formaldehyde gel and transferred to a Hybond-n + Nylon membrane filter (Amersham). The probe was labeled with [α-³²P]dCTP using the Megaprime Random Primer DNA Labeling Kit (Amersham). Hybridization was performed at 65°C for 3 h in rapid hybridization buffer (Amersham) with the labeled probe. The filters were washed and visualized by the BAS-2000 (Fuji Imaging System, Tokyo, Japan).

Methylation Analysis.

Bisulfite sequencing was performed as described with minor modifications (5) . Genomic DNA was extracted using the DNA extractor WB kit (Wako Pure Chemical Industries, Osaka, Japan). After denaturing with NaOH (0.3 m), genomic DNA (2 μg) was incubated with sodium bisulfite (3.1 m) and hydroquinone (0.8 mm; Sigma-Aldrich, St. Louis, MO) (pH 5.0) at 55°C for 16 h, purified, and treated with NaOH (0.2 m) for 10 min at 37°C. Modified DNA (100 ng) was subjected to PCR to amplify the modified promoter sequence of TSLC1 with primers 5′-GTGAGTGACGGAAATTTGTAATGTTTGGTT-3′ and 5′-AATCTAACTTCTTATACACCTTTATTAAAA-3′. The amplification products were subjected to the sequencing in at least three clones to obtain average methylation levels. The criterions for hypermethylation and partial methylation of CpG sites were met when >50% and <50%, respectively, of the PCR products contained bisulfite-resistant cytosines.

Restoration of TSLC1 Expression by a Demethylating Agent.

Cells were seeded at a density of 10⁵ cells onto 10-cm plates at day 0, treated with 10 μm of 5-aza-2′-deoxycytidine (Nacalai Tesque, Kyoto, Japan) for 72 h from day 2 to 5, and harvested at day 7. RT-PCR was performed to confirm the TSLC1 expression.

Transfections.

An ESCC cell line KYSE520 was stably transfected with pcTSLC1 or the empty vector control [pcDNA3.1-Hygro(+)] using LipofectAMINE 2000 reagent (Invitrogen), and cell clones were selected against 200 μg/ml hygromycin (Nacalai Tesque).

Western Blot Analysis.

Cells were washed with PBS and treated with a lysis buffer [50 mm Tris-HCl (pH 7.5), 150 mm NaCl, 5 mm EDTA, and 1% Triton X-100] and protease inhibitor mixture (Roche Diagnostics, Mannheim, Germany) on ice for 15 min and then centrifuged. The protein content was measured using a bicinchoninic acid protein assay reagent (Pierce, Rockford, MA). Cell lysates (20 μg) were electrophoresed on 2–15% gradient polyacrylamide gel (Daiichi Pure Chemicals, Tokyo, Japan) and transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA) with a semidry transfer blot system. After blocking with Tris-buffered saline containing 1% Tween 20 and 5% skim milk for 1 h, the filters were incubated with primary antibody for 1 h, washed, and then incubated with horseradish peroxidase-labeled antirabbit or antimouse IgG (Zymed, San Francisco, CA) as a secondary antibody. Proteins were detected using Western Blotting Luminol Reagent (Santa Cruz Biotechnology, San Diego, CA).

Immunofluorescence Staining.

Cells were seeded on a collagen I-coated coverslip and incubated for 24 h. After being washed with PBS, cells were fixed with 4% paraformaldehyde for 15 min and treated with 0.2% Triton X-100 in PBS for 5 min. Cells were subsequently incubated with a blocking solution (5% BSA in PBS) for 1 h and incubated for 2 h with rabbit anti-TSLC1 polyclonal antibody CC2 as a primary antibody at room temperature. The cells were washed and incubated with rhodamine-conjugated antirabbit IgG (Immunotech, Marseille, France) as a secondary antibody at 1 h at room temperature. Samples were washed, mounted in 80% glycerol, and viewed with a phase-contrast microscope (Olympus, Tokyo, Japan).

Cell Growth Assay.

Cells were plated at a density of 2 × 10⁴ cells/well onto six-well plates (Corning, NY) at day 0 and counted every other day. Each experiment was performed in triplicate wells and repeated three times.

Tumor Formation Assay in Nude Mice.

Suspensions of 1 × 10⁶ cells in PBS (0.1 ml) were injected s.c. into two sites on the flanks of 5–6-week-old male BALB/c athymic nu/nu mice (Clea Japan, Tokyo, Japan), and the tumor growth was estimated by the average volume of tumors at six sites. The tumor volume was calculated by the formula 4π/3 × (L/2 × W/2 × H/2; L = length, W = width, and H = height of the tumor). All of the animal experiments were performed in accordance with institutional guidelines.

Cell Cycle Analysis.

Flow cytometry analysis of DNA content was performed to assess the cell cycle phase distribution. Cells were harvested at the 70% confluent stage and fixed in 70% ethanol at −20°C. After washing with PBS, the cells were treated with PBS containing RNase A (100 mg/ml) at 37°C for 30 min. After centrifugation, the cells were resuspended in PBS containing propidium iodide (50 mg/ml) and stained at room temperature for 30 min. DNA content was evaluated using the FACSCalibur HG and the software CELLQuest version 3.1 (Becton Dickinson Immunocytometer Systems, San Jose, CA) for the histograms. Each experiment was repeated three times.

Cell Migration Assay.

The motility was determined by a micropore chamber assay. Cells (4 × 10⁵) were seeded into the top chamber of a six-well-size micropore polycarbonate membrane filter with 8-μm pores (Becton Dickinson Labware, Lincoln Park, NJ), and the bottom chamber was filled with RPMI 1640 and Ham’s F12 mixed medium containing 2% fetal bovine serum as a chemoattractant. After 16 h of incubation at 37°C, the membranes were fixed and stained by Diff Quik reagent (International Reagents, Inc., Kobe, Japan), and all of the cells that had migrated through the membrane were counted under a light microscope. Each experiment was performed in triplicate wells and repeated three times.

Cell Invasion Assay.

The invasive capacity was determined by an invasion chamber assay. Cells (3 × 10⁴) were seeded into the bottom chamber of a 24-well-size Matrigel-coated micropore membrane filter with 8-μm pores (Becton Dickinson Labware), and the bottom chamber was filled with RPMI 1640 and Ham’s F12 mixed medium containing 10% fetal bovine serum as a chemoattractant. After 22 h of incubation at 37°C, the membranes were fixed and stained by Diff Quik reagent. Then all of the cells that had invaded through the membrane were counted under a light microscope. Each experiment was performed in triplicate wells and repeated three times.

Statistical Analysis.

Fisher’s exact test or Pearson’s χ² test of equality was used to compare clinicopathological backgrounds. The univariate survival analysis was calculated by the Kaplan-Meier method and analyzed by the Log-rank test. Multivariate analysis was estimated by the Cox proportional hazard model. The Tukey-Kramer multiple comparison test was used for evaluation of malignant phenotypes of each stable cell line. The software StatView for Windows version 5.0 (SAS institute, Cary, NC) was used for the analysis. A P < 0.05 was considered statistically significant.

RESULTS

Frequent Loss of TSLC1 Expression in Cell Lines and Primary Tumors from ESCC.

To examine TSLC1 expression in ESCC, we performed RT-PCR in 36 ESCC cell lines and 56 tumor samples. As shown in Fig. 1 ⇓ , we found that 27 of the 36 cell lines (75%) had undetectable levels of TSLC1 expression, whereas 9 of the 36 cell lines (25%) expressed significant amounts of TSLC1. We also found that 28 of the 56 primary tumors (50%) had undetectable levels of TSLC1 expression, whereas 28 of the 56 tumors (50%) expressed TSCL1. In contrast, TSLC1 expression was observed in all of the normal esophageal epithelium and its derived cell lines. Sequencing of the amplified DNA verified the authenticity of the PCR products (data not shown).

Clinicopathological Significance of TSLC1 in ESCC.

To evaluate the relationship of TSLC1 expression with the clinicopathological factors of the ESCC patients, 56 primary tumors examined for TSLC1 expression were divided into two subgroups according to the results of RT-PCR. A group of 28 patients with tumors expressing TSLC1 [TSLC (+)] and a group of 28 patients with tumors lacking TSLC1 expression [TSLC1 (−)]. As shown in Table 1 ⇓ , significant differences in pT, pM, and pathological tumor node metastasis were found between the TSLC1 (+) and TSLC1 (−) groups (P = 0.012, 0.036, and 0.002, respectively) by the statistical analyses. In contrast, no significant differences were seen for age, gender, pN, location, histology, lymphatic involvement, or vascular involvement. As shown in Fig. 2 ⇓ , the univariate analysis using the Kaplan-Meier method revealed that the TSLC1 (−) group tended to have a poorer prognosis than the TSLC1 (+) group (P = 0.079).

Next, to examine whether TSLC1 expression is an independent prognostic factor, the multivariate analysis by the Cox hazard model was performed. Gender (male), pT (pT3-pT4), pN (pN1), pM (pM1), histology (poorly differentiated), and loss of TSLC1 expression [TSLC1 (−)] were included among the parameters (Table 2) ⇓ . This analysis demonstrated that pT [RR, 3.77; CI, 0.92–12.3], pN (RR, 3.83; CI, 1.11–13.2), and loss of TSLC1 expression (RR, 2.47; CI, 1.00–6.05, P = 0.049) were independent prognostic factors; however, pM (RR, 0.88; CI, 0.31–2.55) and histology (RR, 1.2; CI, 0.41–2.98) were not independent prognostic factors.

Silencing of the TSLC1 Gene by the Promoter Methylation.

To investigate the methylation status of the TSLC1 promoter, we directly examined the methylation status of six cytosine residues of CpG sites in a putative promoter sequence upstream from the translation initiation site by bisulfite sequencing in four cell lines, including KYSE270, which expressed TSLC1, and KYSE410, KYSE520, and KYSE960, which did not express it. As shown in Fig. 3 ⇓ A, all of the cytosine residues in KYSE270 DNA were unmethylated, whereas all of the six cytosine residues in KYSE520 DNA and five residues in KYSE410 and KYSE960 DNA were methylated. Especially, the cytosine residues in KYSE520 DNA were all hypermethylated. To examine whether demethylating agents restored the TSLC1 expression, we treated these cell lines with a demethylating agent, 5-aza-2′-deoxycytidine, and examined the TSLC1 expression by RT-PCR (Fig. 3B) ⇓ . We found that the TSLC1 mRNA expressions in KYSE410, KYSE520, and KYSE960 were restored, whereas that of KYSE270 did not change when cells were treated with 5-aza-2′-deoxycytidine.

Growth Suppression of KYSE520 Cells by TSLC1 Both in Vitro and in Vivo.

Inactivation of the TSLC1 gene may confer a variety of malignant phenotypes to ESCC cells. To investigate this possibility, KYSE 520, an ESCC cell line lacking TSLC1 expression because of the promoter hypermethylation, was stably transfected with pcTSLC1, and two independent cell clones TS1 and TS2 expressing TSLC1 were obtained. A mock-transfected clone was also obtained after transfecting an empty vector into KYSE520. TS1 and TS2 showed almost the same levels of TSLC1 mRNA as that detected by Northern blotting (Fig. 4A, a) ⇓ . On the other hand, Western blotting revealed that TS2 expressed higher amounts of TSLC1 protein than did TS1 (Fig. 4A, b) ⇓ . Morphologically, the parental KYSE520 cells, mock-transfected cells, and TS1 cells grew as single cells, whereas TS2 cells were tightly aggregated to each other (Fig. 4B) ⇓ . To determine the subcellular location of TSLC1 protein, we performed immunofluorescence staining. In single cells, TSLC1 protein was expressed mainly in the cytoplasm, but at the beginning of cell adhesion, it relocated to the cell membrane of the attachment site as an interdigitated structure in both TS1 and TS2 cells. In a confluent stage, the protein was located at the cell–cell boundary in irregular forms (Fig. 4C) ⇓ .

To investigate whether TSLC1 expression affects the growth of KYSE520 cells, we performed a cell growth assay. As shown in Fig. 5A ⇓ , we found that TS1 and TS2 grew significantly slower than the parental KYSE520 or mock-transfected cells (P < 0.01). Furthermore, TS2 grew significantly slower than TS1 (P < 0.05). To examine the possible activity of TSLC1 as a tumor suppressor in vivo, s.c. tumor formation assays in nude mice were performed. As shown in Fig. 5, B ⇓ and C, tumor growth was greatly reduced when exogenous TSLC1 was stably expressed in KYSE520 cells. The tumor volumes of the mice at day 28 were suppressed significantly by 64% in TS1 and 98% in TS2 compared with that of mock-transfected cells (P < 0.01). Moreover, suppression of tumor volumes of TS2 was significantly stronger than that of TS1 (P < 0.01). All of the tumors were stained with H&E, and the existence of cancer cells was confirmed (data not shown).

To investigate the growth suppression by exogenous TSLC1 expression in detail, cell cycle changes in TS1, TS2, as well as the parental KYSE520 and mock-transfected cells were examined by flow cytometry. As shown in Fig. 5D ⇓ , the populations of the cells in G₁ phase were significantly higher in TS1 (56.8 ± 1.5%) and TS2 (68.8 ± 1.6%) than that in the mock-transfected cells (48.5 ± 0.8%; P < 0.01).

Suppression of Motility and Invasion of KYSE520 Cells by TSLC1.

As shown above, the clinicopathological analyses of primary ESCC tumors suggest that loss of TSLC1 expression is associated with tumor invasion and metastasis. Therefore, to examine the motility and invasive capacity of the TSLC1-transfected cells, we performed cell migration and cell invasion assays. As shown in Fig. 6 ⇓ A, the migration assay indicated that the motility was significantly suppressed by 70% in TS1 and 80% in TS2 compared with that in mock-transfected cells (P < 0.01). As shown in Fig. 6 ⇓ B, the cell invasion assay revealed that the invasive capacity was also significantly suppressed by 65% in TS1 and 94% in TS2 compared with that in mock-transfected cells (P < 0.01). The invasive capacity of TS2 was suppressed significantly more than that of TS1 (P < 0.01), but no significant differences between TS1 and TS2 were observed in the motility assay.

DISCUSSION

In the present study, we examined the biological and clinical significance of TSLC1 in ESCC. We found that the TSLC1 expression was frequently lost in ESCC cell lines and primary tumors but observed a significant expression in normal esophageal epithelium and its derived cells. This result suggests that TSLC1 may be inactivated during the carcinogenesis of ESCC.

The clinicopathological analyses suggest that loss of TSLC1 expression is associated with the prognosis of ESCC patients and that it can be a useful prognostic marker. It is noteworthy that Yen et al. (4) reported previously that the allelic loss of the distal part of 11q, including 11q23.2, was a significant prognostic factor in a univariate survival analysis. Because TSLC1 is located on 11q 23.2, their findings might well correspond to our results in the present study.

In vitro and in vivo growth suppression by TSLC1 expression was demonstrated in this study. In previous studies, obvious growth suppression in vitro was not reported, and mechanisms of growth suppression were not presented. Here, we examined the cell cycle profile by flow cytometry. A G₁ arrest was observed in TSLC1-transfected ESCC cells but not in mock-transfected cells. Thus, this result suggests that growth suppression through TSLC1 would be caused, at least in part, by a cell cycle arrest at the G₁ phase, in analogy with the actions of other tumor suppressors.

In the clinicopathological analyses of primary ESCC tumors, we have demonstrated that TSLC1 expression is associated with the depth of invasion and status of metastasis, suggesting that TSLC1 is involved in invasion and metastasis of ESCC. Corresponding to this finding, TSLC1 expression strongly suppressed motility and invasion of ESCC cells in in vitro models. As Yageta et al. (12) reported previously, TSLC1 was associated with actin rearrangements induced by 12-O-tetra-decanoylphorbol-13-acetate and suppressed liver metastasis from the spleen of an NSCLC cell line. Masuda et al. (6) reported that TSLC1 protein was involved in cell–cell adhesion using Madin-Darby canine kidney cells. In this study, we also demonstrated that the TSLC1-transfected ESCC cells showed aggregated morphology, where TSLC1 protein accumulated in interdigitated structures at cell–cell membranes. These findings suggest that the TSLC1 protein also is associated with cell–cell adhesion in ESCC. As alterations of E-cadherin also are known to be involved in progression of cancers, including ESCC (17 , 18) , loss of TSLC1 expression may lead ESCC cells to invade or metastasize through disruption of cell–cell interactions.

In vitro cell growth, motility, invasion, and in vivo tumor formation were significantly suppressed in both TS1 and TS2, the stable clones expressing TSLC1. These results indicate that TSLC1 expression suppresses multiple phenotypes related to malignancy of ESCC. Interestingly, TSLC1 mRNA in TS1 and TS2 were detected at the same level by Northern blotting, but the amount of TSLC1 protein detected by Western blotting was much lesser in TS1 than in TS2. These data suggest that the post-transcriptional control might participate in the expression of TSLC1 protein. In this connection, it is noteworthy that in vitro growth, invasion, and in vivo tumor formation are suppressed more dramatically in TS2 than in TS1, suggesting that the degree of suppression of these phenotypes would be dependent on the expression level of TSLC1 protein. On the other hand, no significant difference in the cell motility was observed between TS1 and TS2, implying that the motility of ESCC cells may be affected by a very low amount of TSLC1.

Two-hit inactivation of TSLC1 by promoter hypermethylation was reported in several primary tumors (5 , 8 , 9) . Our study shows a good correlation between loss of TSLC1 expression and promoter methylation of TSLC1. Moreover, restoration of TSLC1 expression by a demethylating agent was observed in the cell lines containing the methylated promoter. These results suggest that promoter methylation is also involved in the inactivation of the TSLC1 gene in ESCC. In prostate cancer, Fukuhara et al. (8) reported that promoter hypermethylation of TSLC1 was observed not only in advanced tumors but also in a subset of relatively early stage tumors. In contrast, loss of TSLC1 expression was preferentially observed in tumors with pT2-pT4 rather than in those with pT1 in ESCC, indicating that inactivation of TSLC1 is a relatively late stage event in the carcinogenesis of ESCC. However, portions of ESCC tumors with pT1 had already lost TSLC1 expression, suggesting that TSLC1 also might be involved in a subset of early ESCC tumors. Additional examination of TSLC1 expression in a series of precancerous tissue samples of the esophagus, such as basal cell hyperplasia, dysplasia, or carcinoma in situ, would be required to elucidate the significance of TSLC1 alteration in the multistage carcinogenesis of ESCC. On the basis of the functional evidence of the involvement of TSLC1 in the suppression of tumor formation of ESCC cell lines, it might be expected that growth of human ESCC tumors could be controlled by restoring TSLC1 expression using demethylating agents.

In conclusion, we have demonstrated that loss of TSLC1 expression correlated with the depth of invasion and the status of metastasis and provided an independent prognostic factor in ESCC. Furthermore, we have clarified that TSLC1 expression modulated cell growth, partly mediated by cell cycle arrest at G₁, and the malignant phenotypes of an ESCC cell. These findings strongly suggest that TSLC1 plays an important role in ESCC progression and would provide a novel molecular target for the treatment of ESCC.