Amplification and Overexpression of CTTN ( EMS 1 ) Contribute to the Metastasis of Esophageal Squamous Cell Carcinoma by Promoting Cell Migration and Anoikis Resistance

Gain of chromosome 11q13 is a common event in esophageal squamous cell carcinoma (ESCC). The cortactin gene (CTTN, also EMS1), located at 11q13, plays a pivotal role in coupling membrane dynamics to cortical actin assembly. This gene has been implicated in the motility of several types of cells. In the present study, we found that the amplification and overexpression of the CTTN gene was associated with lymph node metastasis in ESCC. Functional analysis by small interfering RNA–mediated silencing of CTTN revealed that in addition to the effect on cell migration, CTTN influenced cell invasiveness by anoikis resistance. In vivo assay showed that inhibition of CTTN expression also decreased tumor growth and lung metastasis of ESCC cells. At the molecular level, we showed for the first time that the protective role of CTTN in anoikis resistance was correlated with the activation of the phosphatidylinositol 3-kinase/Akt pathway. Overall, the data suggest that CTTN is an oncogene in the 11q13 amplicon and exerts functions on tumor metastasis in ESCC. (Cancer Res 2006; 66(24): 11690-9)


Introduction
Esophageal cancer is the eighth most common cancer worldwide (1), and esophageal squamous cell carcinoma (ESCC) is the most prevalent type in China.Multiple genetic changes have been found in ESCC, but little is known about major oncogenes and tumor suppressor genes involved in the tumorigenesis of the disease.
Proto-oncogenes are frequently activated by genomic amplification with consequent overexpression, which play an important role in the development of human cancers.Characterization of genes with copy number increase and overexpression in tumor tissues will facilitate the identification of tumor-specific oncogenes.
Our laboratory and other groups previously identified 11q13 as a frequently amplified region in ESCC by comparative genomic hybridization (CGH; refs.[2][3][4].This chromosome locus has also been reported to be amplified in carcinomas of the head and neck, breast, lung, ovary, bladder, etc. (5), suggesting that the 11q13 amplicon may harbor key gene(s) involved in carcinogenesis regardless of tissue types.Several known oncogenes and/or cancer-related genes, such as CCND1, FGF4, FGF3, CTTN, and PAK1, have been mapped to the 11q13 chromosome region.Studies have shown CCND1 and CTTN are likely to be the candidate oncogenes within this region on the basis that only these two genes have been found to be both amplified and overexpressed in tumors (5).In ESCC, the role of CCND1 has been well elucidated (6)(7)(8)(9)(10)(11). Array CGH exhibited increase of CTTN DNA copy number in esophageal tumors (4,12), but the precise role of CTTN in ESCC tumorigenesis and progression is largely unknown.
CTTN has been reported as an actin-associated scaffolding protein, which binds and activates the actin-related protein 2/3 complex, and thus regulates the branched actin networks in the formation of dynamic cortical actin-associated structures (13,14).CTTN amplification and overexpression has been found in breast cancer, bladder cancer, hepatocellular carcinoma, and head and neck squamous cell carcinoma (15)(16)(17)(18)(19).In the present study, we found that amplification and overexpression of CTTN was correlated with lymph node metastasis in esophageal tumor tissues.Cellular function experiment showed that decreased expression of CTTN by RNA interference (RNAi) impaired migration capacity and anoikis resistance of esophageal cancer cells.

Materials and Methods
Tissue specimens.Fresh tissues containing ESCCs and adjacent histologic normal epithelia were procured from surgical resection specimens collected by the Department of Pathology in Cancer Hospital, Chinese Academy of Medical Sciences, Beijing, China.Primary tumor regions and the corresponding histologic normal esophageal mucosa from the same patients were separated by experienced pathologists and immediately stored at À70jC until use.All the patients received no treatment before surgery and signed informed consent forms for sample collection.
DNA extraction and real-time PCR.Tissues were digested with 50 mg/ mL proteinase K in 1% SDS for 12 hours at 48jC and extracted with phenol/ chloroform.DNA was precipitated with cold ethanol and dissolved in water.Genomic copy number of CTTN was evaluated by real-time PCR on an ABI Prism 7000 (Applied Biosystems, Foster City, CA) with SYBR Green PCR core reagents, according to the manufacturer's protocol.GAPDH was applied as the input reference.The primers used for genomic CTTN detection were 5 ¶-TCATTGCTCATTGTGGTAAAGC-3 ¶ as forward and 5 ¶-GTGGGCGGTGTGTCTTTC-3 ¶ as reverse, which amplify the 132-bp uniSTS RH46594 within the CTTN gene.The mean CTTN level of the three real-time quantitative PCR experiments was calculated for each case.Results of the real-time PCR data are presented as C T , which is defined as the threshold PCR cycle number at which an amplified product is first detected.The average C T was calculated for both CTTN and GAPDH, and DC T was determined as the mean of the triplicate C T s for CTTN minus the mean of the triplicate C T s for GAPDH.The DDC T represents the difference between the paired tissue samples, as calculated by the formula DDC T = (DC T of tumor À DC T of normal).The relative copy number of CTTN for a tumor sample compared with its normal epithelial counterpart was expressed as 2 ÀDDC T Â 2. Using this method, the data are presented as the fold change in the target gene (CTTN) in tumor normalized to an internal control gene (GAPDH) and relative to the non-tumor control.
Fluorescence in situ hybridization.Tumor tissues were scrapped and then digested with 0.2% type II collagenase (Sigma, St. Louis, MO).Cells were incubated in 0.075 mol/L KCl hypotonic buffer at 37jC for 60 minutes and fixed in methanol/acetic acid (3:1, v/v) at 4jC.Single-cell suspensions were dropped on cool wet slides.After air-drying overnight, slides were sequentially treated with RNase and pepsin.Denaturation was in 70% formamide, 2 Â SSC (pH 7) for 3 minutes at 75jC.BAC clones of CTTN (RP11-120p20) and 11q23 locus (RP11-94c16) were labeled by random primer technique with cy3-dUTP and FITC-dUTP, respectively.The probe mixture was denatured at 74jC for 8 minutes and then hybridized to denatured slides at 37jC for 48 hours.Post-hybridization washes were carried out in 50% formamide for 15 minutes and twice in 2 Â SSC.Slides were counterstained with 4,6-diamidino-2-phenylindole (DAPI).Gray images were captured with a cooled charged-coupled device camera (Princeton, Inc., Princeton, NJ) equipped with an Opton fluorescence microscope.The images were analyzed using the MetaMorph Imaging System (Universal Imaging Corp., West Chester, PA).
RNA extraction and reverse transcription-PCR.Total RNA was prepared using Trizol reagent (Life Technologies, Gaithersburg, MD) according to the manufacturer's instruction; 5 Ag of total RNA was used to synthesize the first strand of cDNA using SuperScript II RT 200 units/AL (Invitrogen, San Diego, CA).Primers designed for CTTN were 5 ¶-GCGT-CAACCTTTGAGGATGT-3 ¶ as forward and 5 ¶-CTCCTCCAGCTTCCTCCTG-3 ¶ as reverse.GAPDH was used to ascertain the equal amount of cDNA in each reaction.The PCR program was as follows: denaturation at 95jC for 4 minutes; 25 cycles of 95jC for 30 seconds, 60jC for 30 seconds, and 72jC for 40 seconds; followed by a 72jC elongation step for 6 minutes; 10 AL of each PCR product were resolved by 2% (w/v) agarose gel electrophoresis.
ESCC tissue microarray and immunohistochemical staining.A total of 125 formalin-fixed, paraffin-embedded esophageal tumors and the corresponding normal epithelia were placed on the tissue microarray.For each case, normal tissues were repeated thrice, and the cancer tissues were repeated five times.For immunohistochemical analysis, the slides were deparaffinized, rehydrated, then immersed in 3% hydrogen peroxide solution for 10 minutes; heated in citrate buffer (pH 6) at 95jC for 25 minutes; and cooled at room temperature for 60 minutes.The slides were blocked by 10% normal goat serum at 37jC for 30 minutes and then incubated with rabbit polyclonal antibody against CTTN (Santa Cruz  Biotechnology, Santa Cruz, CA) at a dilution of 1:200 or CCND1 (Santa Cruz Biotechnology) at a dilution of 1:100 for 3 hours at 37jC.After washed with PBS, the slides were incubated with biotinylated secondary antibody (diluted 1:100) for 30 minutes at 37jC, followed by streptavidin-peroxidase (1:100 dilution) incubation at 37jC for 30 minutes.Immunolabeling was visualized with a mixture of 3,3 ¶-diaminobenzidine solution.Counterstaining was carried out with hematoxylin.Expression score was determined by staining intensity and immunoreactive cell percentage.Tissues with no staining were rated as 0, with a faint staining or moderate to strong staining in V25% of cells as 1, with moderate staining or strong staining in 25% to 50% of cells as 2, strong staining in z50% of cells as 3.A delta score of tumor tissues minus matched normal epithelia was used to evaluate the extent of overexpression (delta score V0 as negative, 1 as positive, z2 as strongly positive).
Cell culture, transfection, and generation of CTTN-RNAi esophageal cancer cells.The EC9706 was established and studied previously in our laboratory (21).Cells were cultured in RPMI 1640 (Invitrogen) supple-mented with 10% fetal bovine serum (FBS).Cell transfections were done using LipofectAMINE 2000 (Invitrogen) according to the manufacturer's instruction.In the stable transfection, cells were selected with 200 Ag/mL G418, and clones were isolated by serial dilution.Three CTTN RNAi clones (C1, C6, and C19) and three scramble RNAi clones (S4, S5, and S10) were chosen for the subsequent experiments.In the transient transfection with chemical-synthesized siRNA, 60 nmol/L CTTN siRNA or non-silencing siRNA were used, and cells were harvested 72 hours after transfections.
Wound-healing assay.Cells were grown to confluence and then wounded using a yellow pipette tip.Three wounds were made for each sample, and all were photographed at the zero time point and at subsequent time points.Assays were repeated thrice in each clone.
Haptotactic migration and Matrigel chemoinvasion assays.For the haptotactic cell migration assay, 1 Â 10 5 CTTN RNAi, scramble RNAi, and parental EC9706 cells were seeded on a fibronectin-coated polycarbonate membrane insert (6.5 mm in diameter with 8.0-Am pores) in a Transwell apparatus (Costar, Cambridge, MA) and maintained in RPMI 1640.RPMI 1640 containing 20% FBS was added to the lower chamber.After incubation for 12 hours at 37jC in a CO 2 incubator, the insert was washed with PBS, and cells on the top surface of the insert were removed by wiping with a cotton swab.For the Matrigel chemoinvasion assay, the procedure was similar with the haptotactic cell migration assay, except that the Transwell membrane was coated with 300 ng/AL Matrigel (BD Biosciences, San Jose, CA), and the cells were incubated for 24 hours at 37jC.Cells that migrated to the bottom surface of the insert were fixed with methanol and stained by 0.4% crystal violet and then subjected to microscopic inspection.Cells were counted based on five field digital images taken randomly at Â200.
Soft agar assay for colony formation.The procedure was based on a modified method as described (22).CTTN RNAi, scramble RNAi, and parental EC9706 cells were plated in six-well plates at a density of 500 per well.For each clone, three independent wells were examined.After 3 weeks of incubation at 37jC, 5% CO 2 , colonies were stained with 0.2% p-iodonitrotetrazolium violet and counted.
Xenograft assays in nude mice.The mixture of C1, C6, and C19 or S4, S5, and S10 were injected into female nude mice (Vital River, Beijing, China) via s.c. or tail vein injection.Ten mice were s.c.injected and eight were tail vein injected with 1 Â 10 6 cells per animal for each group.The mice that survived the procedure were sacrificed 4 weeks after injection and examined for s.c.tumor growth or metastases development.To clearly observe the metastasis nodules in the lung, we fixed the lung in Bouin's solution and counted the visually observable metastases.Then the tissues were embedded in paraffin, and the sections were stained with H&E.For the survival analysis, 1 Â 10 6 cells were injected via the tail vein per animal, and eight mice were injected for each group.
Terminal deoxynucleotidyl transferase-mediated nick-end labeling assay.The formalin-fixed, paraffin-embedded sections of s.c.tumors from nude mice were analyzed by using the Dead-end Fluorometric Terminal Deoxynucleotidyl Transferase-Mediated Nick-End Labeling (TUNEL) System (Promega).The assay was done according to the user manual.Total DNA was stained with DAPI, and apoptotic cells appeared in green color with the FITC filter under a fluorescent microscope.
Assessment of anoikis and the inhibitor treatment.Cells were prevented from adhering to the plastic dishes by culturing them in dishes coated with PolyHEMA (Sigma) as described previously (23).After 24 hours of growth in suspension, cells were harvested for apoptosis measurement using Annexin V-FITC apoptosis detection kit (Sigma) and subsequently analyzed by flow cytometry.
Statistical analysis.We statistically evaluated experimental results using the Kruskal-Wallis test, Mann-Whitney test, and ANOVA test.All tests of significance were set at P < 0.05.For survival analysis, Kaplan-Meier survival curves were constructed, and differences between them were tested by the log-rank test.

Results
Amplification of CTTN in ESCC cancer tissues.The copy number of genomic CTTN was evaluated by real-time PCR of RH46594, an intragenic locus of CTTN in 20 cases (tumors and matched adjacent histologic normal epithelia).GAPDH was applied as the internal reference (Fig. 1).Melting curves were done to confirm the PCR specificity (Supplementary Fig. S1).Increased DNA dosage was observed in 14 of 20 (70%) tumors.Positive correlation was found between CTTN amplification and lymph node metastases (Table 1).
CTTN amplification was then found in 6 of another 10 esophageal cancer tissues by fluorescence in situ hybridization (FISH) on interphase nuclei (Supplementary Fig. S2A).Signals of the control 11q23 probe indicated that the increase in CTTNhybridizing loci did not result from polyploidy of chromosome 11.
As CTTN is known to be frequently coamplified with CCND1, the same samples were also subjected to CCND1 DNA copy number detection by real-time PCR.Increased genomic DNA dosage was observed in 14 of 20 (70%) tumors for WI-16756, an intragenic locus of CCND1 (Supplementary Fig. S3A and B).Statistical analysis showed that CCND1 amplification was not associated with lymph node metastasis (Table 1).As these results were from 20 cases, a larger patient number is needed to further verify the conclusion.
Overexpression of CTTN mRNA and protein in ESCC.CTTN mRNA expression was measured by reverse transcription-PCR (RT-PCR) in primary esophageal tumors and matched normal adjacent epithelia.Eight tumors of the 16 cases showed elevated CTTN mRNA level (Fig. 2A).Western blot showed that CTTN protein was up-regulated to different degrees in four of six malignant tissues (Fig. 2B).
Immunohistochemical analysis was done using CTTN antibody on the tissue microarray.Normal epithelia showed negative or weak immunoreactions, but 86 of 125 cases exhibited CTTNpositive or strongly positive staining (Fig. 2C).Statistical analysis indicated that there were significant correlations between CTTN expression and lymph node metastasis and stage (Table 2).
CCND1 expression was also examined on tissue microarray.In the 117 of 125 cases that could be analyzed, CCND1 overexpression was detected in 55 tumors by immunohistochemical staining (data not shown).Statistical analysis displayed a positive correlation between CCND1 and CTTN protein expression (Spearman rho = 0.359; P < 0.01).Nevertheless, CCND1 overexpression did not correlate with lymph node metastasis (Table 2).RNAi of CTTN in EC9706 cells.EC9706 is an ESCC cell line with CTTN and CCND1 amplification (Supplementary Figs.S2B and  S3C).Two pairs of oligonucleotides targeting sequences in the coding region of the CTTN gene, CTTN RNAi-1 and RNAi-2 were synthesized as short hairpin-interfering RNAs and inserted into the pSilencer 3.1-H1 neo siRNA expression vector.Both sequences were present in the transcripts for the two known isoforms of CTTN.Then we transiently transfected EC9706 with these vectors and examined the mRNA level of CTTN 48 hours after transfection.RT-PCR exhibited that CTTN RNAi-1 could not suppress the expression of CTTN effectively, whereas CTTN RNAi-2 worked well (Fig. 3A).Thus, CTTN RNAi-2 was selected to stably transfect EC9706 cells.As shown in Fig. 3A, CTTN RNAi-2 markedly suppressed the expression of CTTN protein in C1, C6, and C19.Parallel experiments were carried out using the scramble RNAi clones (S4, S5, and S10), which did not cause a reduction of CTTN expression.In the following description, we referred to CTTN RNAi-2 as CTTN RNAi.
Reduced expression of CTTN decreased cell motility in EC9706.Wound-healing assays showed that CTTN RNAi cells migrated much more slowly than scramble RNAi and parental EC9706 cells.Twenty hours after wounding, the scramble RNAi and parental EC9706 cells became confluent, whereas the CTTN RNAi cells migrated only two-cell to four-cell distance (Fig. 3B).When cell motility was examined using the haptotactic cell migration assay and Matrigel chemoinvasion assay, 5-to 8-fold more parental EC9706 and scramble RNAi cells migrated to the bottom chamber than did the CTTN RNAi cells (Fig. 3C and D).These experiments had been repeated thrice in each clone, and the results were similar.
As the apparent CTTN-induced increase in migration could be the result of an increase in the adhesion of tumor cells to the substrate, we evaluated the adhesive abilities of these three

Cancer Research
Cancer Res 2006; 66: (24).December 15, 2006 groups of cells by measuring the number of cells attached to Matrigel.No significant difference was detected by MTS assays among the groups (data not shown).Thus, knockdown of CTTN expression by RNAi drastically suppressed the mobility of EC9706 cells in vitro.
Suppressed CTTN expression in EC9706 cells inhibited colony formation in soft agar and s.c.tumor growth in nude mice.To investigate the effect of CTTN expression on esophageal tumor growth, we examined the in vitro cell growth rates by MTS cell proliferation assays.No significant difference was detected among CTTN RNAi, scramble RNAi, and parental EC9706 cells (data not shown).However, colony-forming activity of CTTN RNAi cells under anchorage-independent condition was markedly decreased comparing with scramble RNAi and parental EC9706 cells in soft agar assay (Fig. 4A).
Then we examined the effect of CTTN on tumor growth in vivo.CTTN RNAi, scramble RNAi, and parental EC9706 cells were inoculated s.c.into nude mice.All the three groups of cells developed tumors, but the tumors formed by scramble RNAi or parental EC9706 cells grew faster than the CTTN RNAi ones.When the mice were sacrificed 4 weeks after injection, the tumor weights of scramble RNAi or parental EC9706 cells were significantly heavier than those of the tumors formed by CTTN RNAi (P < 0.01; Fig. 4B).CTTN expression was validated in the formalin-fixed, paraffin-embedded sections of the s.c.tumors by immunohistochemical staining (Fig. 4C).
In view that tumor growth is determined by the balance of cell proliferation and programmed cell death, cell proliferation-related protein Ki-67 was analyzed on s.c.tumor sections by immunohistochemistry.No difference was observed between among three groups of tumors (data not shown).Then apoptosis in the s.c.tumor sections was examined by TUNEL assays.Very few apoptotic cells were found in the tumors derived from scramble RNAi and parental EC9706 cells.In contrast, significantly more apoptotic cells were detected in the CTTN RNAi tumors (Fig. 4D).These results indicated that CTTN RNAi in esophageal cancer cells suppressed in vivo tumor growth by promoting apoptosis.
RNAi of CTTN promoted anoikis in EC9706 cells via inactivation of PI3K/Akt pathway.Cell cycle progression and apoptosis with or without induction of UV exposure and doxorubicin treatment of CTTN RNAi, scramble RNAi, and parental EC9706 cells were analyzed by propidium iodide staining using flow cytometry.CTTN down-regulation in EC9706 cells did not affect these phenotypes (data not shown).
Considering that CTTN overexpression enhanced lymphoid metastasis, and anoikis is one type of apoptosis involved in tumor metastasis (24), we tested the detachment-induced apoptosis.Of the stable transfected cells, total percentage (early and late apoptosis) of apoptotic CTTN RNAi cells was about thrice as that of the control and parental cells in the assessment by Annexin V-FITC/propidium iodide staining (Fig. 5A).Repeated experiment data were shown in Table 3.
To explore the molecular mechanism by which CTTN protected EC9706 cells from anoikis, we examined the potential effect of CTTN on the activation of several known pathway associated with anoikis.Cells cultured on polyHEMA-coated dishes were treated with PI3K inhibitor LY294002 or MEK inhibitor PD98059.As shown in Fig. 5A and Table 3, both LY294002 and PD98059 abrogated the survival protection in scramble RNAi and parental EC9706 cells.Then the activation of PI3K/Akt, MEK/Erk, and EGFR signaling was measured by Akt phosphorylation on Ser 473 , Erk phosphorylation on Tyr 204 , and EGFR expression, respectively.Only pAkt was down-regulated in CTTN siRNA cells.No difference in Erk activation or EGFR expression was observed in these three groups of cells (Fig. 5B), suggesting that they were not downstream targets accounting for the protective effect against anoikis of CTTN.
We also evaluated anoikis and related pathways in transient transfected EC9706 cells with another chemical-synthesized CTTN siRNA, and the results were similar with the stable clones (Fig. 5C; Table 4).Thus, our data suggested that CTTN promoted cell survival under anchorage-independent conditions via the PI3K/Akt pathway.
CTTN down-regulation in EC9706 cells decreased lung metastases in nude mice and prolonged survival time of tumor-bearing mice.As increased cell migration and suppression of anoikis both contributes to metastatic potential of cancer cells, it is worthy to investigate whether inhibition of CTTN in the highly metastatic EC9706 cells would affect metastasis ability.CTTN RNAi, scramble RNAi, and parental EC9706 cells were introduced via tail vein into nude mice.Four weeks after injection, the animals were anatomized to examine metastatic growth.Of the eight mice in each group, six mice inoculated with scramble RNAi cells, and five with parental EC9706 cells, but none with CTTN RNAi cells developed visually observable lung nodules.Representative lungs and their corresponding H&E staining were shown in Fig. 6A, and the number of visually observable lung nodules was plotted in Fig. 6B.Only small tumor nodules were found under microscope in H&E sections of two lungs derived from CTTN RNAi cells.No visible or microscopy metastases were detected in other organs of the mice.
In survival analysis, mice injected with scramble RNAi cells and parental EC9706 via tail vein resulted in more rapid pulmonary metastases and death than those with CTTN RNAi cells (P < 0.01; Fig. 6C).These data suggested that CTTN expression strongly promoted tumor metastases in vivo.
Figure 5. Down-regulation of CTTN promoted anoikis in EC9706 cells via PI3K/Akt pathway.A, flow cytometry analysis of apoptotic cells by Annexin V-FITC/propidium iodide staining.Plots for attached cells are in the first row.In the anoikis assay, cells were cultured on polyHEMA-coated dishes for 24 hours.For the third and fourth rows, cells were treated with 20 Amol/L PI3K inhibitor LY294002 and 40 Amol/L MEK inhibitor PD98059, respectively, when undergoing anoikis.B, cell lysates of stable CTTN RNAi, scramble RNAi, and parental EC9706 cells were immunoblotted with antibodies against pAkt on Ser 473 , total Akt, pErk on Tyr 204 , total Erk, EGFR, and h-actin.C, cell lysates of CTTN siRNA and non-silencing siRNA transient transfected cells and parental EC9706 were immunoblotted for CTTN, pAkt, total Akt, pErk, total Erk, EGFR, and h-actin.Relative ratios of absorbance for CTTN to h-actin, pAkt to total Akt, pErk to total Erk, EGFR to h-actin were plotted.Columns, mean from triplicate experiments; bars, SD. *, P < 0.01, compared with non-silencing and parental EC9706 cells.
We subsequently detected CTTN at the genomic DNA and protein levels.Although different lots of samples were used for genomic DNA, mRNA, and protein analysis in our study, the frequency of CTTN amplification by real-time PCR (70%) and by FISH (60%) in tumors was parallel to that of overexpression by RT-PCR examination (50%) and immunohistochemical staining (68.8%).These data suggested that CTTN amplification is probably responsible for overexpression in most cases.Further statistical analysis revealed positive correlation between CTTN amplification/ overexpression and lymph node metastasis in ESCC.This is accordant with previous observations that CTTN or 11q13 amplification was associated with lymphoid metastasis in breast, head and neck, larynx, and oral cancers (27)(28)(29)(30).However, it needs to be clarified whether both CTTN and CCND1 or only one of these two genes is the prediction factor for lymph node metastasis in ESCC, for they are frequently coamplified on 11q13.There are contradictory reports about the correlation between CCND1 amplification and expression with lymph node metastasis in ESCC (7,(31)(32)(33)(34).Our experiments indicated neither amplification nor overexpression of CCND1 was associated with lymph node metastasis, whereas our real-time PCR result was from a small patient number.The immunohistochemical data derived from 117 cases could better support the conclusion that CTTN is an independent marker at 11q13 for ESCC lymphoid metastasis.In addition, incidence rate of CCND1 amplification (14 of 20, 70%) was higher than that of CCND1 overexpression (55 of 117, 47%) in our study.This situation has been described in esophageal carcinoma by two different groups (35,36).They used the same samples for FISH and immunohistochemical analysis, and the frequencies of CCND1 copy number increase were 32.5% and 18% more than that of overexpression, respectively.Other mechanisms may influence CCND1 protein level besides amplification, such as alterations in the synthesis or stability of the CCND1 protein and regulation by transcription factors.
The cellular role of CTTN is related to membrane dynamics and cortical actin assembly, including cell migration, morphogenesis, adhesion, receptor-mediated endocytosis and pathogen invasion, etc. (37).More recently, functional analysis in breast cancer, hepatocellular carcinoma, and head and neck squamous cell carcinoma has revealed that CTTN promotes invasiveness of cancer cells (19,22,38,39).In the present study, CTTN RNAi in EC9706 cells impaired the migration capacity.Tail vein injection of CTTN RNAi cells decreased lung metastases in nude mice and prolonged survival time compared with the controls.These observations confirm that CTTN serves a metastatic function in ESCC.
Given the complexity of the metastasis process, cancer cells must acquire a series of traits that enable them to overcome multiple barriers erected by normal tissues.Enhancement in migration abilities is advantageous to tumor invasion, which is the principal mechanism reported to account for the role of CTTN in tumor metastasis (14).Additionally, CTTN ectopic expression potentiates bone metastasis of breast cancer cells by increasing the adhesive affinity for bone marrow endothelial cells (22).Therefore, CTTN overexpression endows cancer cells with various capabilities for metastasis.
Loss of adhesion is a stress that metastatic cancer cells encounter en route.Resistance to anoikis may allow survival of cancer cells during systemic circulation and facilitate secondary tumor formation in distant organs.Notably, decreased CTTN expression in EC9706 cells was associated with increased anoikis.To our knowledge, this is the first report that CTTN was involved in  cell anoikis, which extends our understanding of the role CTTN plays in tumor metastasis.It is reasonable to presume that the reduction in lung metastasis due to knockdown of CTTN after tail vein injection was relevant with the impaired capacity of anchorage-independent growth, besides the decreased cell motility.
In addition, when cells were injected s.c., they should suffer from inadequate cell-matrix contact.Prevention of cancer cell apoptosis favored the establishment of tumor colonies.Cancer cells colonizing in foreign tissues also needed to confront the inappropriate microenvironment when expanding and invading to adjacent tissues, anoikis-resistant capabilities would remain helpful.In view that CTTN did not affect cell proliferation in vitro, and TUNEL assays of s.c.tumors in nude mice indicated the involvement of apoptotic mechanisms, we speculated that the different growth rates of s.c.tumors in nude mice derived from CTTN RNAi, scramble RNAi, and parental EC9706 cells were due to the disparity in the protecting ability from anoikis.Molecular mechanisms in anoikis resistance have been described concerning several signal pathways in different cell types.PI3K/Akt, MEK/Erk, and EGFR are important pathways that mediate survival signals in detachment-induced apoptosis (24,40).In the present study, PI3K inhibitor promoted anoikis, and Akt phosphorylation lowered after CTTN RNAi.Although MEK inhibitor induced anoikis, activated Erk level did not alter after CTTN suppression.As overexpressed EGFR allow cells survive from anoikis (40), and CTTN inhibits ligand-induced down-regulation of EGFR (41), we detected EGFR expression after CTTN RNAi.Giving that our results were obtained in the conditions that all cells were maintained in 10% FBS, inhibition of EGF-induced EGFR down-regulation by CTTN did not take place in our experiment.Thus, PI3K/Akt, not MEK/Erk or EGFR signaling, contributed to CTTN-related survival in detached cells.However, when PI3K activity was inhibited by LY294002, CTTN RNAi could still increase anoikis in our experiment.Other pathways, in addition to PI3K, may participate in CTTN-related anoikis resistance.
It remains unclear how CTTN associates with PI3K/Akt pathway.One possibility is that CTTN directly interacts with PI3K, activates PI3K, and subsequently makes Akt phosphorylated.The binding of CTTN to PI3K SH2 domain has been revealed by in vitro assays (42).An alternative to consider is that CTTN activates PI3K via filamentous-actin (F-actin).The recruitment of cytosolic PI3K to the membrane, which determines its activity, is partly dependent on F-actin polymerization (43,44).CTTN is an important regulator of actin rearrangements (14).Thus, the effect of CTTN on PI3K/Akt signaling may be mediated through F-actin.Accumulating evidences have implicated that cytoskeletal alterations are potentially involved in anoikis (24), but the mechanisms are still not completely understood.Because CTTN is a scaffold protein regulating cortical actin assembly, our result that exhibits a role of CTTN in anoikis resistance provides new insights into the influence of cytoskeletal organization on the survival pathway in anoikis.
Taken together, we show that both CTTN amplification and overexpression are correlated with lymph node metastasis in ESCC.Our in vitro and in vivo data indicate that CTTN contributes to tumor aggressiveness through increasing the capacity of cell migration and anoikis resistance.These results suggest that CTTN is an oncogene involved in the progression of ESCC.

Figure 1 .
Figure 1.Real-time PCR detection of intragenic loci of CTTN and GAPDH in matched esophageal tumors and normal epithelia.Thick horizontal lines represent the threshold for C T estimation.CTTN copy number was higher in the tumor (black curve ) than the normal epithelial tissue (gray curve ) compared with GAPDH .

Figure 2 .
Figure 2. Increased expression of CTTN mRNA and protein in esophageal cancer.A, up-regulated CTTN mRNA level was detected in 8 of 16 tumors (T ) compared with normal adjacent epithelia (N ) by RT-PCR.B, CTTN protein level was increased in four malignant tissues by Western blot analysis.C, example case that CTTN is overexpressed in esophageal tumors by immunohistochemical staining on the tissue microarray.There were three normal tissues and five cancer tissues in each case.Bottom, magnification of tissues in black frames.

Figure 3 .
Figure 3. Effects of CTTN RNAi on cell migration and invasion.A, analysis of CTTN expression in EC9706 cells.CTTN mRNA level was examined by RT-PCR in transient transfected CTTN RNAi-1, CTTN RNAi-2, scramble RNAi, and parental EC9706 cells.C1, C6, and C19 stably expressing CTTN RNAi-2 exhibited down-regulation of CTTN by Western blot.B, CTTN-deficient cells migrate more slowly in wound-healing assays.A confluent monolayer of cells was wounded and photographed at the indicated time points.Representative results of three independent experiments.C and D, representative photos and statistical plots of haptotactic migration assay (C ) and Matrigel chemoinvasion assay (D) in CTTN RNAi, scramble RNAi, and parental EC9706 cells.The number of CTTN RNAi cells that transversed the Transwell membranes in the haptotactic migration assay and Matrigel chemoinvasion assay was significantly different from that of the scramble RNAi and parental EC9706 cells.Columns, mean of three individual experiments in each clone; bars, SD.

Figure 4 .
Figure 4. RNAi-mediated inhibition of CTTN in EC9706-affected colony formation in soft agar and s.c.tumor growth in nude mice.A, representative photos and statistic plot for each clone in soft agar assay.B, tumors shown 28 days after injection.Tumor weights were plotted.C, formalin-fixed, paraffin-embedded sections of the s.c.tumors from CTTN RNAi, scramble RNAi, and parental EC9706 cells were analyzed by immunohistochemical staining of CTTN.D, s.c.tumor sections were analyzed by TUNEL assay.Total DNA were stained with DAPI, and apoptotic cells were visualized by TUNEL with FITC staining.Percentage of TUNEL-positive cells in s.c.tumor sections was plotted.Columns, mean; bars, SD.

Table 1 .
Gene amplification and tumor clinicopathologic features

Table 3 .
Apoptotic cell percent by flow cytometry analysis in stable transfected cells and parental EC9706

Table 4 .
Apoptotic cell percent by flow cytometry analysis in transient transfected cells and parental EC9706