TSLC1 Is a Tumor Suppressor Gene Associated with Metastasis in Nasopharyngeal Carcinoma

Research Article

Hong Lok Lung,1 Arthur Kwok Leung Cheung,1 Dan Xie,2,4 Yue Cheng,1,5 Fung Mei Kwong,1 Yoshinori Murakami,6 Xin-Yuan Guan,2 Jonathan Shuntong Sham,2 Daniel Chua,2 Alexey I. Protopopov,7 Eugene R. Zabarovsky,7 Sai Wah Tsao,3 Eric J. Stanbridge,8 and Maria Li Lung1

1Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Kowloon, Hong Kong (SAR), People’s Republic of China; The Departments of 2Clinical Oncology and 3Anatomy, University of Hong Kong, Pokfulam, Hong Kong (SAR), People’s Republic of China; 4State Key Laboratory of Oncology in Southern China, Cancer Center, Sun Yat-Sen University, Guangzhou, People’s Republic of China; 5Genetics Branch, National Cancer Institute, National Naval Medical Center, Bethesda, Maryland; 6Tumor Suppression and Functional Genomics Project, National Cancer Center Research Institute, Tokyo, Japan; 7Microbiology and Tumor Biology Center, Center for Genomics and Bioinformatics, Karolinska Institute, Stockholm, Sweden; and 8Department of Microbiology and Molecular Genetics, University of California, Irvine, California

Abstract 0.44 Mb region was located around the markers D11S1300 and In up to 87% of nasopharyngeal carcinoma (NPC) clinical D11S1391. In a third region, high allelic loss was also observed at tumor specimens, there was either down-regulation or loss of marker D11S4484, where a tumor suppressor gene, tumor TSLC1 TSLC1 TSLC1 gene expression.Using a tissue microarray and suppressor in lung cancer 1 ( ), is located (1). The immunohistochemical staining, the frequency of down-regu- gene was originally identified as a tumor suppressor in non–small- lated or loss of expression of TSLC1 in metastatic lymph node cell lung cancer by combinatorial analyses of yeast artificial NPC was 83% and the frequency of loss of expression of TSLC1 chromosome transfer into human non–small-cell lung cancer cells was 35%, which was significantly higher than that in primary with a tumorigenicity assay in nude mice (2). NPC (12%).To examine the possible growth-suppressive The gene expression analysis revealed absence or low TSLC1 activity of TSLC1 in NPC, three NPC cell lines, HONE1, expression levels of mRNAin four highly tumorigenic HNE1, and CNE2, were transfected with the wild-type TSLC1 NPC cell lines, HONE1, HK1, HNE1, and CNE1. In addition, the TSLC1 gene cloned into the pCR3.1 expression vector; a reduction of methylation study results show that the promoter region colony formation ability was observed for all three cell lines.A was hypermethylated in all four NPC cell lines and reexpression ¶ tetracycline-inducible expression vector, pETE-Bsd, was also of the gene occurs in HONE1 cells after 5-aza-2 -deoxycytidine used to obtain stable transfectants of TSLC1.There was a treatment (1). Hence, the mode of silencing of this candidate dramatic difference between colony formation ability in the tumor suppressor gene in NPC is consistent with promoter TSLC1 presence or absence of doxycycline when the gene is shut off hypermethylation. Mutation of the gene in primary NPC or expressed, respectively, with the tetracycline-inducible has been reported, but is not common, whereas promoter f system.Tumorigenicity assay results show that the activation methylation of this gene is detected in 36% of primary tumors TSLC1 of TSLC1 suppresses tumor formation in nude mice and and 60% of NPC cell lines (3). promoter hypermethylation functional inactivation of this gene is observed in all the is also reported in other cancers, including non–small-cell tumors derived from tumorigenic transfectants.Further lung (2), prostate (4), breast (5), esophageal (6), and cervical studies indicate that expression of TSLC1 inhibits HONE1 cell cancers (7). TSLC1 growth in vitro by arresting cells in G -G phase in normal Besides the epidemiology studies of in various cancers, 0 1 TSLC1 culture conditions, whereas in the absence of serum, TSLC1 to date, evidence for functional involvement of in tumor induced apoptosis.These findings suggest that TSLC1 is a suppression has been reported in very few studies including tumor suppressor gene in NPC, which is significantly non–small-cell lung, prostate, esophageal, and cervical cancers associated with lymph node metastases. (Cancer Res 2006; (2, 4, 6, 7). In the present study, we focus on the functional TSLC1 66(19): 9385-92) role of in the tumorigenesis of NPC. To examine the possible growth-suppressive activity of TSLC1 in NPC, HONE1, HNE1, and CNE2 cells were transfected with the wild-type TSLC1 Introduction gene and the colony formation assay was done. ApCR3.1 In a previous study, we identified three nasopharyngeal expression vector and a tetracycline-inducible expression vector, carcinoma (NPC) critical regions in 11q22-23. One critical region pETE-Bsd, were used in the TSLC1 functional studies. By using of 0.36 Mb was mapped near the marker D11S2000 and a second this technology of inducible gene expression, stable transfectants of essential genes in cancer cells such as TSLC1 can be obtained. An in vivo tumorigenicity assay for the TSLC1 transfectants was done. Astably transfected tumor-suppressive TSLC1 cell line was in vitro Requests for reprints: Maria Li Lung, Department of Biology, Hong Kong further examined by growth, cell cycle, and apoptosis University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), analyses. In addition, the clinical relevance of TSLC1 in NPC was People’s Republic of China. Phone: 852-2358-7307; Fax: 852-2358-1559; E-mail: examined in NPC biopsies and in tissue microarray studies by [email protected]. I2006 American Association for Cancer Research. reverse transcription-PCR (RT-PCR) and immunohistochemistry doi:10.1158/0008-5472.CAN-06-0590 approaches. www.aacrjournals.org 9385 Cancer Res 2006; 66: (19). October 1, 2006

Materials and Methods in the membrane, then the expression of TSLC1 was considered normal. Down-regulated expression of TSLC1 was noted when the percentage of Cell lines and culture conditions. The recipient NPC HONE1 cell line, positive cells ranged from 20% to 70%. Loss of expression of TSLC1 was donor chromosome 11 mouse hybrid cell line (MCH556.15), and two NPC determined when <20% of tumor cells were positively stained. Cases with cell lines, HNE1 (8) and CNE2 (9), were maintained as previously described <50 tumor cells were excluded from statistical analysis. (1). The immortalized nasopharyngeal epithelial cell line NP69 was cultured Transient transfection and colony formation assay. The full-length as described (10). Construction of a pETE-Bsd responsive vector and a wild-type TSLC1 cDNA(2) was ligated to the pCR3.1 neomycin-resistance trans HONE1 cell line, HONE1-2, producing the tetracycline -activator tTA, tagged expression plasmid. The pCR3.1-TSLC1 recombinant and control TSLC1 was described by Protopopov et al. (11). Stable transfectants with pCR3.1 (Invitrogen) vector-alone plasmids were transfected into the NPC transgene or with pETE-Bsd vector alone were maintained in culture cell lines with Lipofectamine 2000 Reagent (Invitrogen). Atotal of 4 Â 105 A A medium containing 500 g/mL neomycin and 5 g/mL blasticidin. cells were seeded in six-well plates and were transfected with 1 Agof TSLC1 Tissue specimens. For gene expression analysis, 38 primary NPC plasmid DNAfor 5 hours. The cells were subsequently split into 100-mm biopsies were collected from Queen Mary Hospital in 2001. The Hong Kong dishes. After 18 days of selection in DMEM/10% FCS containing 500 Ag/mL patients were of ages 30 to 78 years (mean, 51 years), with a male-to-female neomycin, colonies were fixed and stained with Giemsa to assess the ratio of 1.9:1. The tumor specimens encompassed 14 primary NPC cases transfection efficiency. without metastasis, 24 lymph node metastatic NPC, and 2 with distant For inducible expression of TSLC1, the full-length TSLC1 cDNAwas metastasis. They were classified as undifferentiated to poorly differentiated subcloned into the pETE-Bsd responsive vector containing a selectable squamous cell carcinomas. For preparation of the NPC tissue microarray, 80 blasticidin-resistance gene. The recombinant pETE-Bsd-TSLC1 and control NPC specimens and 20 specimens of nasopharyngeal mucosa from non- pETE-Bsd vector were transfected into the HONE1-2 cells as described NPC diseases were collected in the Cancer Institute of Sun Yat-Sen above. After 18 days of selection in DMEM/10% FCS containing 5 Ag/mL University, as previously described (12). These included 50 primary NPC blasticidin and 500 Ag/mL neomycin with or without 0.5 Ag/mL doxycycline cases without metastasis and 30 lymph node metastatic NPC tissues, all of (a tetracycline analogue), colonies were fixed and stained. which were poorly differentiated squamous cell carcinomas. The Guang- Stable transfection of TSLC1. To generate stable clones, which express zhou patients were of ages 14 to 72 years (mean, 47 years), with a male-to- wild-type TSLC1, HONE1-2 was transfected with pETE-Bsd-TSLC1 and 0.5 female ratio of 1.3:1. Ag/mL doxycycline was added to the HONE1-2 cells before and after RT-PCR. Total RNAwas extracted from cells with the RNeasy Midi Kit transfection. Independent single clones were obtained after 2 to 3 weeks of (Qiagen, Valencia, CA). One microgram of total RNA was reverse drug selection post-transfection. Both total RNAand protein were extracted transcribed with 200 units of SuperScriptII reverse transcriptase (Invitro- from the clones and RT-PCR, real-time RT-PCR, and Western blot analyses gen, Carlsbad, CA). For PCR amplification of specific cDNA derived from were used to screen for positive clones. The pETE-Bsd vector alone was TSLC1 gene, a pair of primers was used as previously described (4). For transfected to HONE1-2 cells as described above and those stable clones GAPDH amplification of glyceraldehyde-3-phosphate dehydrogenase ( ), the served as controls. ¶ ¶ ¶ primer sequences were 5 -GAGTCAACGGATTTGGTCGT-3 and 5 -ATCCA- In vivo tumorigenicity assay. The tumorigenicity of each cell line was ¶ CAGTCTTCTGGGTGG-3 . PCR was done in a GeneAmp PCR System 9700 assayed by s.c. injection as described by Cheng et al. (15). In brief, 1 Â 107 programmable thermal controller. cells were injected into 4- to 8-week-old female athymic BALB/c nu/nu Quantitative real-time RT-PCR analysis. Total RNAwas reverse mice. Atotal of six sites were tested for each cell line and tumor growth in transcribed as described. Real-time PCR was done in a MX3000P machine animals was checked weekly. If tumor formation was observed, then (Stratagene, La Jolla, CA) with TaqMan PCR core reagent kit (Applied representative tumors were reconstituted into cell culture for subsequent TSLC1 Biosystems, Foster City, CA), -specific primers and probe (Applied Bio- molecular analyses. For inhibition of the tetracycline-inducible expression systems). Human a-tubulin primer and probe reagents (Applied Biosystems) of TSLC1 in vivo, 200 Ag/mL doxycycline was added to the drinking water of were used as the normalization control in subsequent quantitative analysis. mice 1 week before injection and the water with doxycycline was changed Transcript quantification was done in triplicate for every sample and twice a week. reported relative to a-tubulin. Cell growth assay. The cell growth of the TSLC1 tranfectant and the Western blot analysis. Preparation of cell lysates, protein fractionation, vector-alone clones was analyzed with the 3-(4,5-dimethylthiazol-2-yl)-2,5- and transfer were as previously described (12). After transferring to diphenyl-tetrazolium bromide (MTT) assay (16). In brief, 2 Â 103 cells were polyvinylidene difluoride membranes (pore size, 0.45 Am; Millipore, seeded on day 0 and cell growth was measured daily until day 7. Avolume of Billerica, MA), the membranes were blocked with 5% skim milk and 30 AL of MTT (Sigma Chemical Co., St. Louis, MO) solution (5 mg/mL) was primary antibody incubation was done with C-16 (1:2,500; Santa Cruz then added and cells were further incubated at 37jC for 2 hours. The Biotechnology, Santa Cruz, CA) for the TSLC1 protein and with Ab-1 absorbance at 540 nm was determined with a Opsys MR microplate reader (1:5,000; Calbiochem, Darmstadt, Germany) for a-tubulin. The signals were (Thermo Labsystems, Beverly, MA). Measurements of cell growth by MTT visualized by enhanced chemiluminescence method according to the assay were expressed as percent inhibition according to the following for- instructions of the manufacturer (Amersham, Uppsala, Sweden). mula: % inhibition = [(absorbance of vector-alone transfectant À absorbance Tissue microarray and immunohistochemical staining. The tissue of TSLC1 transfectant) / absorbance of vector-alone transfectant] Â 100%. microarray was constructed according to the method previously described Fluorescence-activated cell sorting analysis. The cell cycle distribu- (12, 13). Immunohistochemistry studies were done with the standard tion and apoptosis of TSLC-C54 and BSD-C1 were analyzed by FACScan streptavidin-biotin-peroxidase complex method as described (13). The flow cytometry as described by Ito et al. (6). In brief, 1 Â 106 cells were tissue microarray slides were incubated with rabbit anti-TSLC1 polyclonal seeded and allowed to attach to the bottom of a T25 culture flask overnight. antibody (CC2; 1:250 dilution) overnight at 4jC. The slides were then Both the floating and adherent cells were collected, washed with PBS, and incubated with a biotinylated goat anti-rabbit serum for 30 minutes and fixed with cold 70% ethanol. The fixed cells were then treated with 1 unit of subsequently reacted with a streptavidin-peroxidase conjugate and 3¶,3¶- DNase-free RNase and incubated for 30 minutes at 37jC. Propidium iodide diaminobenzidine. Known immunostaining-positive slides for lung cancer (1 mg/mL; Sigma Chemical) was added directly to the cell suspension and a were used as positive controls. total of 10,000 fixed cells were analyzed by FACScan (Becton Dickinson, For evaluation of the staining, the nonmalignant and malignant tissues Franklin Lakes, NJ). were scored for TSLC1 by assessing the site of positive staining in cell Cell viability assay. Briefly, 3 Â 105 cells were plated into 35-mm- membranes (Fig. 1B). The membranous immunostaining of TSLC1 was diameter culture dishes and allowed to grow overnight. The number of observed mainly in the epithelial tissue with no staining in the stroma. viable cells was determined with a hemocytometer using the trypan blue Patients were classified into three groups according to the grading method dye exclusion assay (17). The viable cells were visible as clear cells whereas previously described (14). When >70% of tumor cells were positively stained the dead cells were stained blue.

Cancer Res 2006; 66: (19). October 1, 2006 9386 www.aacrjournals.org

4,6-Diamidineo-2-phenylindole stain apoptosis assay. Apoptosis was cance of TSLC1 in NPC, we studied the gene expression of TSLC1 in observed with the DNA-specific stain 4,6-diamidineo-2-phenylindole (DAPI) 38 NPC specimens by RT-PCR. After 38 cycles of PCR amplification, to observe nuclear condensation (18). Cells were fixed and incubated for 15 33 of 38 (87%) primary tumors showed no or decreased TSLC1 j A minutes at 37 C with 2.5 g/mL DAPI (Sigma Chemical), washed with PBS, expression when compared with the immortalized normal and observed under a fluorescence microscope. Fluorescence signals were nasopharyngeal epithelial cell line NP69 and human chromosome captured using SPOT software (Diagnostic Instruments, Sterling Heights, A MI) on an Olympus BX51 microscope (Tokyo, Japan). The apoptotic cells 11 mouse hybrid cell line MCH556.15 (Fig. 1 ). However, there was TSLC1 were counted and photographed. Atotal of 20 fields per group were no significant association of the decreased gene expression m2 counted and the results were represented as the average percentage of with pathologic staging, histology, sex, and age using the test. apoptotic cells out of total cells per field. The small sample size might preclude meaningful statistical Statistical analysis. The m2 and Fisher’s exact test were used for analysis analysis to correlate the level of TSLC1 expression with clinical of significant differences in TSLC1 expression level detected by tissue variables for these specimens. microarray between primary and metastatic NPC. Differences were The 100-sample NPC tissue microarray was analyzed for P in vitro considered statistically significant for < 0.05. All other assay expression of TSLC1 protein by immunohistochemistry. Informa- F results represent the arithmetic mean SE of triplicate determinations of tive tissue microarray cases were observed in 75 of the 100 cases. at least two independent experiments done under the same conditions. The noninformative samples included lost or unrepresentative Student’s t test was used to determine the confidence limits in group comparison and P < 0.05 was considered as statistically significant. samples and samples with too few tumor cells or with inappro- priate staining. In 20 cases of normal nasopharyngeal mucosa, all the 8 informative cases showed normal expression of TSLC1 Results protein. However, in 67 informative NPC cases, 43% (29 of 67) cases TSLC1 expression in NPC biopsy specimens and NPC tissue were observed with down-regulated expression of TSLC1 and 21% microarray. To determine the involvement and clinical signifi- (14 of 67) were detected with loss of expression of this protein.

Figure 1. A, RT-PCR analysis of TSLC1 gene expression analysis of NPC specimens. The immortalized normal nasopharyngeal epithelial cell line NP69 and normal chromosome 11 donor cell line MCH556.15 were used as positive controls. The DNA ladder (L) indicates the sizes of the PCR bands, which are shown on the right (bp). GAPDH served as an internal control. B, immunohistochemical staining of TSLC1 in NPC tissue microarray containing normal nasopharyngeal mucosa, primary NPC, and lymph node metastatic NPC. Representative staining results of TSLC1 expressed in normal nasopharyngeal mucosa and tumor tissues showing loss of TSLC1 expression (Â400 magnification). www.aacrjournals.org 9387 Cancer Res 2006; 66: (19). October 1, 2006

Table 1. Expression of TSLC1 in primary and lymph node transfected and vector alone–transfected HONE1-2 cells. Negative metastatic NPC tissues controls without addition of vector DNAto HONE1-2 cells showed no expression of TSLC1 (Fig. 2B). TSLC1 Normal Down-regulated Loss of Total To further test the inhibitory effect of on colony expression expression expression formation in NPC, HONE1 cells and two other NPC cell lines, HNE1 and CNE2, were transfected with the pCR3.1-TSLC1 Normal 8 (100%) 0 (0%) 0 (0%) 8 recombinant and control pCR3.1 plasmids. Western blot results Primary NPC 19 (46%) 17 (42%) 5 (12%) 41 show that the TSLC1 protein was almost undetectable in these Metastatic NPC 5 (19%) 12 (46%) 9 (35%) 26 three NPC cell lines, and when the protein was expressed in each of P = 0.024* P = 0.11* P = 0.007* the transfected cell lines (Fig. 2B), a significant decrease in number of neomycin-resistant colonies occurred (P = 0.0012, P = 0.02, and P = 0.0089 for HONE1, HNE1, and CNE2, respectively) when they *Comparison between primary and metastatic NPC. were transfected with the wild-type TSLC1 cDNA, as compared with vector alone (Fig. 2A). Stable transfection of TSLC1. Based on the screening by The frequency of lost expression of TSLC1 in lymph node RT-PCR with 30 cycles of PCR amplification, six TSLC1-transfected metastatic NPC was 35%, which was significantly higher than that clones were chosen for further studies: TSLC-C7, TSLC-C16, TSLC- in primary NPC (12%; P < 0.01, m2 test; Table 1). Representative C19, TSLC-C20, TSLC-C38, and TSLC-C54. As can be seen, in the results of the NPC tissue microarray are shown in Fig. 1B. absence of doxycycline when the gene was activated, TSLC1 was Transient transfection of TSLC1 and colony formation overexpressed in all the six clones, whereas in the presence of assay. The gene expression analysis showed that TSLC1 was doxycycline, the TSLC1 expression was suppressed in all six clones consistently down-regulated in HONE1 and other NPC cell lines (1) (Fig. 3A). However, the levels of TSLC1 expression in those and biopsy specimens. To provide functional proof, the growth- suppressed clones were even higher than the positive control cell suppressive activity of TSLC1 was evaluated by the colony line MCH556.15, the vector-alone clone (Bsd-C1), or the recipient formation assay. Atetracycline-inducible transgene system (11) HONE1-2 cells. Amore accurate real-time RT-PCR was applied to was used to study the functional role of TSLC1 in NPC. HONE1-2 quantify the fold changes in the TSLC1 stable transfectants. TSLC1 cells producing tetracycline trans-activator tTAwere used for gene expressions of TSLC-C16, TSLC-C38, and TSLC-C54 are the transfections. In the absence of doxycycline, the TSLC1 protein was most suppressed by doxycycline (Fig. 3B) and were chosen for transiently expressed up to 96 hours (Fig. 2B) and a significant protein analysis. The addition of doxycycline in vitro decreased the reduction (P = 0.0046) in the number of blasticidin-resistant transcription of ectopic TSLC1 of TSLC-C16, TSLC-C38, and TSLC- colonies was observed in HONE1-2 cells transfected with TSLC1 C54 by 750-, 266-, and 323-fold, respectively. However, the levels of cDNA, as compared with vector alone (Fig. 2A). The addition of TSLC1 expression in those three clones were still higher than doxycycline inhibited TSLC1 protein expression. No significant MCH556.15 by 2.2- to 4.7-fold. As shown in Fig. 3C, the protein change of colony number was observed between the TSLC1- expression of TSLC1 in the three chosen clones was induced

Figure 2. Colony formation assay of TSLC1 transfectants. A, HONE1-2 cells were transfected with pETE-Bsd-TSLC1 or pETE-Bsd in the presence or absence of doxycycline (dox) and HONE1, HNE1, and CNE2 cells were transfected with pCR3.1-TSLC1 or pCR3.1. The colony-forming ability was calculated by comparing the number of colonies in TSLC1 transfectants to that of vector alone. Each treatment was done in triplicate. *, P < 0.05, statistically significant difference from vector alone. B, Western blot analysis of the TSLC1-transfected HONE1-2 cells and vector-alone transfectants. HONE1-2 cells were transfected in the absence or presence of doxycycline. Western blotting was done 1 to 4 days after transfection. C, Western blot analysis of HONE1, HNE1, and CNE2 transfected with pCR3.1-TSLC1 or pCR3.1 for 4 days. Transfection without plasmid DNA served as a negative control. a-Tubulin was used as the internal control.

Cancer Res 2006; 66: (19). October 1, 2006 9388 www.aacrjournals.org

Figure 3. Expression analysis and in vivo tumorigenicity of pETE-Bsd-TSLC1 transfectants. A, RT-PCR analysis of TSLC-C7, TSLC-C16, TSLC-C19, TSLC-C20, TSLC-C38, TSLC-C54, and Bsd-C1 in the presence or absence of doxycycline. TSLC1 expression in HONE1-2 cells was used as a negative control and the expression in MCH556.15 served as a positive control. B, real-time PCR analysis of TSLC1 in same cell lines as in (A). Relative expression (yCt) of TSLC1 in each cell line compared with Bsd-C1 was studied. C, Western blot analysis of TSLC1 in the selected transfectants. D, tumor growth kinetics of HONE1-2 cells, BSD-C1, and TSLC-C54 with and without doxycycline treatment and TSLC-C16 and TSLC-C38 without doxycycline in nude mice. Points, average tumor volume of all sites inoculated for each cell population.

without the doxycycline treatment and the enhanced protein TSLC-C54 was consistently slower than BSD-C1 for the 7-day expression was reduced in the presence of doxycycline and the incubation period. This difference gradually increased from day 1, reduced protein levels were slightly higher than the Bsd-C1. Hence, reached maximum (65%) on day 5, and decreased thereafter. To TSLC-C16, TSLC-C38, and TSLC-C54 were selected for the in vivo further study the growth inhibition mechanism of TSLC1, the cell tumorigenicity assay of TSLC1. cycle status of TSLC-C54 and Bsd-C1 was determined by DNAflow In vivo tumorigenicity assay of TSLC1. After injection into cytometry analysis. When the gene was expressed (in the absence nude mice, both the recipient HONE1-2 cells and the vector- of doxycycline), there was an increase in the relative number of alone transfectant Bsd-C1 form palpable tumors in all six sites of cells in G0-G1 phase from 48.3% to 59.6% and a significant decrease injection within 2 to 3 weeks in the presence or absence of of cells in S phase from 13.7% to 3.6% (Fig. 5B). Taken together with doxycycline. Tumorigenicity can be suppressed by overexpressing the MTT assay results, a G0-G1 growth arrest was observed in the TSLC1, as observed with clone TSLC-C54; no appearance of TSLC1 transfectant when compared with the vector-alone clone tumor was observed in all six sites of injection. In the presence after 24 hours of incubation. For TSLC-C54 cultures serum starved of doxycycline, however, there was still a clear growth inhibition for 24 hours, the growth inhibitory function is even more dramatic effect. No tumor suppression was observed for TSLC-C16 and than in the presence of serum; the viable cell count is decreased by TSLC-C38, as all the six injection sites formed tumors within 2 88% when compared with BSD-C1 (P = 0.00048), whereas in the weeks for both clones. The tumor growth kinetics for the TSLC1 presence of serum, it is only decreased by 16% (Fig. 5C). DAPI transfectants, Bsd-C1, and HONE1-2 cells are shown in Fig. 3D. staining the nuclei of this TSLC1 tranfectant showed that there was To check the stability of the newly transferred TSLC1 following a significant increase (3.8-fold) in the percentage of apoptotic cells tumor formation in the two tumorigenic TSLC1 clones, total observed growing under serum starvation (P = 0.0026; Fig. 5D). The RNAand protein were isolated from each tumor and cells from typical features of apoptotic cell death, including plasma one of the tumors (T6) were reconstituted in culture medium. membrane blebbing, nuclear condensation and fragmentation, cell Results of RT-PCR analysis for tumors and tumor segregants shrinkage, and formation of apoptotic bodies, were observed with derived from TSLC-C16 and TSLC-C38 show that the TSLC1 TSLC-C54 (data not shown). The change of percentage of apoptotic gene expression was significantly reduced, as compared with cells in the two clones is not significant when serum was added their corresponding transfectants (Fig. 4A). In concordance with (Fig. 5D). the RT-PCR results, the TSLC1 protein was not detected in all six tumors and their tumor segregants from both TSLC-C16 and TSLC-C38 (Fig. 4B). Discussion In vitro growth-suppressive activities of TSLC1. To determine Our previous genotyping results showed a high frequency of the mechanisms for TSLC1-induced tumor suppression, in vitro cell allelic loss in the tumor segregants derived from HONE1/ growth and cell cycle analysis were done for the tumor-suppressive chromosome 11 hybrids at the TSLC1 locus in the region of transfectant TSLC-C54 and the vector-alone transfectant Bsd-C1. 11q23.2-3. In addition, the gene expression of TSLC1 was shown to Figure 5A shows that the growth rate of the transgene-expressing be down-regulated in the tumorigenic NPC cell lines including www.aacrjournals.org 9389 Cancer Res 2006; 66: (19). October 1, 2006

Figure 4. Loss of expression of TSLC1 in tumors and tumor segregants derived from tumorigenic transfectants. A, RT-PCR analysis of TSLC1 in tumors TSLC-C16-T1 to TSLC-C16-T6 and TSLC-C38-T1 to TSLC-C38-T6 and their representative tumor segregants, TSLC-C16-T6-TS and TSLC-C38-T6-TS. Gene expression of TSLC-C16 and TSLC-C38 served as positive controls. B, Western blot analysis of TSLC1 in the six tumors and a tumor segregant derived from each clone.

HONE1, HK1, CNE1, CNE2, and HNE1 cells. The inactivation of formation ability of HONE1 cells using both the pCR3.1 and the TSLC1 expression in the NPC cell lines was attributed to promoter inducible pETE-Bsd vector systems. The specificity of the growth- hypermethylation (1, 3). Aprevious study showed this epigenetic suppressive effect of TSLC1 was supported by the absence of silencing of TSLC1 in 34% of primary NPCs and no somatic suppression in the presence of doxycycline when the transgene mutations of TSLC1 were detected in the same set of NPC samples was not expressed. Similar growth inhibition effects were observed (3). This suggests that TSLC1 may belong to the class of genes that with other transfected NPC cell lines showing down-regulated predispose to cancer through haploinsufficiency in the hemizygous TSLC1 expression. state and, therefore, will not show a second mutation in the In the current study, we show that the transfectant TSLC-C54 remaining wild-type allele in tumors (19, 20). was unable to induce tumor formation in nude mice when the Based on the low expression of TSLC1 in HONE1 cells, we gene is expressed in the absence of doxycycline. In the presence of transfected wild-type TSLC1 into these cells to obtain functional doxycycline, however, there was still a clear growth inhibition evidence of the involvement of this gene in NPC. The in vitro effect. This may be explained by any leakiness in TSLC1 expression colony formation assay provides functional evidence that intro- because even weak expression of TSLC1 due to such leakage might duction of wild-type TSLC1 is sufficient to inhibit the colony have a strong growth-inhibiting effect, as previously observed for

Figure 5. A, cell growth of TSLC-C54 and Bsd-C1 determined by MTT assay. Columns, percent change in growth inhibition by comparing the absorbance detected in TSLC1 transfectants to that of vector alone. B, fluorescence-activated cell sorting analysis of TSLC-C54 and Bsd-C1. The average percentage of cells in G0-G1, S, and G2-M phases is calculated from three independent experiments. Representative propidium iodide staining results of TSLC-C54 and Bsd-C1. C, total number of viable cells of TSLC-C54 and Bsd-C1 in the presence or absence of serum was determined by the trypan blue dye exclusion assay. *, P < 0.05, statistically significant difference from the vector-alone clone. D, apoptotic cell staining of TSLC-C54 and Bsd-C1 in the presence or absence of serum. The percentage of apoptotic cells was calculated.

Cancer Res 2006; 66: (19). October 1, 2006 9390 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2006 American Association for Cancer Research. Tumor Suppressor Gene TSLC1 in Nasopharyngeal Carcinoma other tumor suppressor genes such as HYA22, RASSF1C, and transition contributes to invasion and metastasis of carcinoma NPRL2, using the currently used tetracycline-inducible system cells from epithelial tumors (29, 30). (21–23). As shown in both RT-PCR and Western blot analyses, the The majority of NPC specimens show down-regulated or lost addition of doxycycline in vitro could not reduce the ectopic TSLC1 gene and protein expression, whereas all normal nasopha- TSLC1 expression of transfectant TSLC-C54 to the basal expression ryngeal mucosa samples showed normal expression of this protein of HONE1 cells or the vector-alone clone Bsd-C1. The leakage in tissue microarray analysis. TSLC1 protein was found primarily in expression of TSLC1 in TSLC-C54 in vitro was even higher than the plasma membrane in the cells of the normal nasopharyngeal that observed in the TSLC1-positive cell line MCH556.15. Moreover, mucosae and this is in agreement with the reported subcellular it is likely that leakage is stronger in vivo than in vitro, as can be localization of TSLC1 (2, 6). The frequency of down-regulated or seen in the current study and as mentioned in previous studies loss of expression of TSLC1 in metastatic lymph node NPC was 83% using the current inducible expression system (21–23). As can be (21 of 26), which is significantly higher than in primary tumors in seen, TSLC-C54 expresses TSLC1 at a higher than normal level the nasopharynx (54%; 22 of 41). This suggests that the inactivation even in the presence of doxycycline and, hence, it is tumor of TSLC1 protein expression might be associated with lymph node suppressive in nude mice with or without the doxycycline metastasis. Due to its significant homology with the neural cell treatment. adhesion molecule and several other immunoglobulin superfamily For the other two transfectants, TSLC-C16 and TSLC-C38, proteins, TSLC1 has been suggested to be involved in cell adhesion tumors formed in all nude mice tested in the absence of (31). Disruption of cell adhesion might be associated with the doxycycline. The subsequent RT-PCR and Western blot analyses invasion or metastasis of cancer cells to adjacent or distal tissues showed that after inoculation of those two clones in nude mice, (32). In the current study, TSLC1 seems to have prognostic signifi- losses of mRNAand protein expressions of TSLC1 were observed in cance in NPC patients and TSLC1 expression may be useful in iden- all six tumors and the representative tumor segregant derived from tifying NPC patients at risk of developing lymph node metastases. both clones. Importantly, it is clear from this functional Clinicopathologic analyses have revealed that the inactivation of inactivation that TSLC1 is capable of antagonizing tumor growth TSLC1 occurs more frequently in tumors in advanced stages. For of HONE1 in nude mice. Furthermore, in the presence of selective example, loss of TSLC1 expression in primary esophageal cancers medium (G418 and Bsd), the majority of cells of tumor segregants was also detected more frequently in tumors at pathologic stages II derived from TSLC-C16 and TSLC-C38 were unable to survive,9 and to IV in comparison with those at stage I (6). In addition, TSLC1 most likely tumors are derived from those G418/Bsd-sensitive cells. promoter hypermethylation in primary non–small-cell lung cancer To date, evidence for functional involvement of TSLC1 in tumor was observed preferentially in tumors at pathologic stages Ib to IV suppression has only been reported in non–small-cell lung, rather than in tumors at stage Ia (33). In an immunohistochemical prostate, and esophageal cancers (2, 4, 6). This is the first report study, Uchino et al. (14) showed that TSLC1 expression was for the functional role of TSLC1 in NPC. inversely correlated with advanced disease stage, lymph node Results of the in vitro growth assay and cell cycle analysis are involvement, lymphatic permeation, and vascular invasion. Fur- consistent with increased TSLC1 expression negatively regulating thermore, 4-year survival and disease-free survival are significantly the G1-to-S phase transition. This tumor-suppressive effect of shorter in patients with lung adenocarcinomas lacking TSLC1 TSLC1 is similar to that observed previously in esophageal and lung expression. Taken together with the current clinical analysis of cancers (6, 24). Another possible mechanism for TSLC1 to induce TSLC1 in NPC, TSLC1 is likely to be responsible for the biological tumor-suppressive activity is through cell apoptosis, as high aggressiveness of tumors, including invasion and metastasis. expression levels of TSLC1 from a recombinant adenovirus vector The current study showed a high frequency down-regulation of inhibited cell proliferation and induced apoptosis in A549 cells and TSLC1 gene expression in NPC specimens. In addition, the strongly suppressed the s.c. tumor growth of A549 cells in nude frequency of down-regulated or loss of expression of TSLC1 in mice (25). In the current study, we also show that in the absence of metastatic lymph node NPC is significantly higher than in primary serum, TSLC1 induced apoptosis, and a similar observation has tumors in the nasopharynx. In addition, the growth-suppressive been reported in the phosphatase and tensin homologue–induced function of TSLC1 in NPC was shown. By using a tetracycline- apoptosis (26). They showed that phosphatase and tensin inducible system, we show that the activation of TSLC1 suppresses homologue inhibits cell proliferation and induces apoptosis by tumor formation in nude mice and functional inactivation of this the down-regulation of insulin-like growth factor-I receptor and it gene is observed in the tumorigenic transfectants. Inhibition of TSLC1 is possible does its tumor-suppressive function in a growth cell growth by G0-G1 phase arrest and induction of apoptosis also factor–dependent manner. contribute to its tumor-suppressive function. These findings The other reported mechanism of the tumor-suppressive activity suggest that TSLC1 is a tumor suppressor gene in NPC, which is of TSLC1 is through the heterophilic trans-interaction with class significantly associated with lymph node metastases. Because of I–restricted T cell–associated molecule to enhance the cytotoxicity the strong association of TSLC1 with metastatic NPC, it may serve of natural killer (NK) cells and the secretion of IFN-g from CD8+ T as a good candidate biomarker associated with tumor metastasis cells to attack the TSLC1-expressing cells. Tumor cells expressing and progression. TSLC1 are efficiently rejected by NK cells in the early stages of inoculation into the peritoneal cavity of nude mice (27). Recently, Masuda et al. (28) reported that TSLC1 suppressed the induction of Acknowledgments epithelial-to-mesenchymal transitions by regulating the activation Received 2/14/2006; revised 6/1/2006; accepted 7/19/2006. of small Rho GTPases. It is known that epithelial-to-mesenchymal Grant support: The Research Grants Council of the Hong Kong Special Adminis- trative Region, People’s Republic of China: Grant number HKUST 6113/01M and CA03/ 04.01 to M.L. Lung and the Swedish Cancer Society, the Swedish Research Council, STINT, the Swedish Institute, the Royal Swedish Academy of Sciences and INTAS to 9 Unpublished observation. E.R. Zabarovsky. www.aacrjournals.org 9391 Cancer Res 2006; 66: (19). October 1, 2006

The costs of publication of this article were defrayed in part by the payment of page We thank Prof. Yan Fang (State Key Laboratory of Oncology in Southern China, charges. This article must therefore be hereby marked advertisement in accordance Cancer Center, Sun Yat-Sen University, Guangzhou, People’s Republic of China) for with 18 U.S.C. Section 1734 solely to indicate this fact. providing us the NPC specimens for tissue microarray analysis.

References affected chromosomal region, 3p21. J Gene Med 2002;4: 22. Li J, Wang F, Protopopov A, et al. Inactivation of 397–406. RASSF1C during in vivo tumor growth identifies it as a 1. Lung HL, Cheng Y, Kumaran MK, et al. Fine mapping 12. Lung HL, Bangarusamy DK, Xie D, et al. THY1 is a tumor suppressor gene. Oncogene 2004;23:5941–9. of the 11q22–23 tumor suppressive region and involve- candidate tumour suppressor gene with decreased 23. Li J, Wang F, Haraldson K, et al. Functional ment of TSLC1 in nasopharyngeal carcinoma. Int J expression in metastatic nasopharyngeal carcinoma. characterization of the candidate tumor suppressor Cancer 2004;112:628–35. Oncogene 2005;24:6525–32. gene NPRL2/G21 located in 3p21.3C. Cancer Res 2004; 2. Kuramochi M, Fukuhara H, Nobukuni T, et al. TSLC1 is 13. Xie D, Sham JS, Zeng WF, et al. Heterogeneous expres- 64:6438–43. a tumor-suppressor gene in human non-small-cell lung sion and association of h-catenin, p16 and c-myc in multi- 24. Sussan TE, Pletcher MT, Murakami Y, Reeves RH. cancer. Nat Genet 2001;27:427–30. stage colorectal tumorigenesis and progression detected by Tumor suppressor in lung cancer 1 (TSLC1) alters 3. Hui AB, Lo KW, Kwong J, et al. Epigenetic inactivation tissue microarray. Int J Cancer 2003;107:896–902. tumorigenic growth properties and gene expression. of TSLC1 gene in nasopharyngeal carcinoma. Mol 14. Uchino K, Ito A, Wakayama T, et al. Clinical Mol Cancer 2005;4:28. Carcinog 2003;38:170–8. implication and prognostic significance of the tumor 25. Mao X, Seidlitz E, Truant R, Hitt M, Ghosh HP. Re- 4. Fukuhara H, Kuramochi M, Fukami T, et al. Promoter suppressor TSLC1 gene detected in adenocarcinoma of expression of TSLC1 in a non-small-cell lung cancer cell methylation of TSLC1 and tumor suppression by its the lung. Cancer 2003;98:1002–7. line induces apoptosis and inhibits tumor growth. gene product in human prostate cancer. Jpn J Cancer 15. Cheng Y, Stanbridge EJ, Kong H, Bengtsson U, Oncogene 2004;23:5632–42. Res 2002;93:605–9. Lerman MI, Lung ML. Afunctional investigation of 26. Zhao H, Dupont J, Yakar S, Karas M, LeRoith D. PTEN 5. Allinen M, Peri L, Kujala S, et al. Analysis of 11q21–24 tumor suppressor gene activities in a nasopharyngeal inhibits cell proliferation and induces apoptosis by loss of heterozygosity candidate target genes in breast carcinoma cell line HONE1 using a monochromosome down-regulating cell surface IGF-IR expression in cancer: indications of TSLC1 promoter hypermethyla- transfer approach. Genes Chromosomes Cancer 2000;28: prostate cancer cells. Oncogene 2004;23:786–94. tion. Genes Chromosomes Cancer 2002;34:384–9. 82–91. 27. Boles KS, Barchet W, Diacovo T, Cella M, Colonna M. 6. Ito T, Shimada Y, Hashimoto Y, et al. Involvement of 16. Hansen MB, Nielsen SE, Berg K. Re-examination and The tumor suppressor TSLC1/NECL-2 triggers NK cell TSLC1 in progression of esophageal squamous cell further development of a precise and rapid dye method and CD8+ T cell responses through the cell surface carcinoma. Cancer Res 2003;63:6320–6. for measuring cell growth/cell kill. J Immunol Methods receptor CRTAM. Blood 2005;106:779–86. 7. Steenbergen RD, Kramer D, Braakhuis BJ, et al. TSLC1 1989;119:203–10. 28. Masuda M, Kikuchi S, Maruyama T, et al. TSLC gene silencing in cervical cancer cell lines and cervical 17. Lung HL, Ip WK, Chen ZY, Mak NK, Leung KN. (tumor suppresor in lung cancer)1 suppresses epithelial neoplasia. J Natl Cancer Inst 2004;96:294–305. Comparative study of the growth-inhibitory and apo- cell scattering and tubulogenesis. J Biol Chem 2005;280: 8. Yao KT, Zhang HY, Zhu HC, et al. Establishment and ptosis-inducing activities of black tea theaflavins and 42164–7. characterization of two epithelial tumor cell lines (HNE- green tea catechin on murine myeloid leukemia cells. 29. Thiery JP. Epithelial-mesenchymal transitions in 1 and HONE-1) latently infected with Epstein-Barr virus Int J Mol Med 2004;13:465–71. tumour progression. Nat Rev Cancer 2002;2:442–54. and derived from nasopharyngeal carcinomas. Int J 18. Clay CE, Namen AM, Atsumi G, et al. Influence of J 30. Gotzmann J, Mikula M, Eger A, et al. Molecular Cancer 1990;45:83–9. series prostaglandins on apoptosis and tumorigenesis of aspects of epithelial cell plasticity: implications for 9. Lerman MI, Sakai A, Yao KT, Colburn NH. DNA breast cancer cells. Carcinogenesis 1999;20:1905–11. local tumor invasion and metastasis. Mutat Res 2004; sequences in human nasopharyngeal carcinoma cells 19. Fero ML, Randel E, Gurley KE, Roberts JM, Kemp CJ. 566:9–20. that specify susceptibility to tumor promoter-induced The murine gene p27Kip1 is haplo-insufficient for 31. Masuda M, Yageta M, Fukuhara H, et al. The tumor neoplastic transformation. Carcinogenesis 1987;8:121–7. tumour suppression. Nature 1998;396:177–80. suppressor protein TSLC1 is involved in cell-cell 10. Lo AK, Liu Y, Wang XH, et al. Alterations of biologic 20. Tang B, Bottinger EP, Jakowlew SB, et al. Trans- adhesion. J Biol Chem 2002;277:31014–9. properties and gene expression in nasopharyngeal forming growth factor-h1 is a new form of tumor 32. Murakami Y. Functional cloning of a tumor suppres- epithelial cells by the Epstein-Barr virus-encoded latent suppressor with true haploid insufficiency. Nat Med sor gene, TSLC1, in human non-small cell lung cancer. membrane protein 1. Lab Invest 2003;83:697–709. 1998;4:802–7. Oncogene 2002;21:6936–48. 11. Protopopov AI, Li J, Winberg G, et al. Human cell 21. Kashuba VI, Li J, Wang F, et al. RBSP3 (HYA22) is a 33. Fukami F, Fukuhara H, Kuramochi M, et al. Promoter lines engineered for tetracycline-regulated expression of tumor suppressor gene implicated in major epithelial methylation of the TSLC1 gene in advanced lung tumors tumor suppressor candidate genes from a frequently malignancies. Proc Natl Acad Sci U S A 2004;101:4906–11. and various cancer cell lines. Int J Cancer 2003;107:53–9.

Cancer Res 2006; 66: (19). October 1, 2006 9392 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2006 American Association for Cancer Research. TSLC1 Is a Tumor Suppressor Gene Associated with Metastasis in Nasopharyngeal Carcinoma

Hong Lok Lung, Arthur Kwok Leung Cheung, Dan Xie, et al.

Cancer Res 2006;66:9385-9392.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/66/19/9385

Cited articles This article cites 33 articles, 6 of which you can access for free at: http://cancerres.aacrjournals.org/content/66/19/9385.full#ref-list-1

Citing articles This article has been cited by 13 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/66/19/9385.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/66/19/9385. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.