Eighty pairs of clinical specimens were obtained from patients with HCC who were hospitalized in the First Affiliated Hospital of Nanjing Medical University with informed consent. Adjacent non-tumor tissues were excised 2 cm from the edge of the primary focus. Both HCC specimens and adjacent non-tumor tissues were immediately stored in liquid nitrogen after excision and confirmed by pathological examination. The protocols for investigations involving humans and animals were approved by the Institutional Animal Care and Use Committee at Nanjing Medical University.

Human hepatocellular carcinoma cell lines (QGY-7703, Focus, Hep3B, HepG2, HepG2.2.15, Huh7, LM3, LM6, MHCC-H, MHCC-L, PLC/PRF/5, SK-hep-s, SNU-398, WRL-68, and YY-8103) obtained from the Chinese National Human Genome Center at Shanghai were used in this study. All cell lines were propagated at 37°C in a 5% CO₂ humidified incubator in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, penicillin (50 U/ml), and streptomycin (50 μg/ml).

Total RNA was extracted from clinical samples or cell lines using TRIzol solution (Invitrogen) according to the manufacturer’s protocol and reverse-transcribed into cDNA using a M-MLV reverse transcriptase kit (Promega). Primers used in semiquantitative RT-PCR and quantitative real-time PCR were as follows: COTE1, 5′-GGGCTCTGACCTAGGCTTCT-3′ (forward) and 5′-ACAGAAGCTCTCCCAGTCCA-3′ (reverse); β-actin (loading control): 5′-AGAGCCTCGCCTTTGCCGATCC-3′ (forward) and 5′-CTGGGCCTCGTCGCCCACATA-3′ (reverse). All primers were synthesized by Shanghai Biosune Co. Ltd.

HCC cells grown on polylysine-treated slides were washed twice with phosphate-buffered saline (PBS), fixed with 4% paraformaldehyde on ice for 30 min, and blocked with 5% BSA at room temperature. Cells were stained with primary antibody [goat anti-COTE1 antibody (1:50) or mouse anti-WWOX antibody (1:50), Santa Cruz Biotechnology, CA, USA] at 4°C overnight, followed by incubation with secondary antibody [Cy5-conjugated anti-mouse secondary antibody (1:200, red); Cy3-conjugated anti-goat antibody (1:200, green), Molecular Probes Inc., Eugene, OR, USA] at room temperature for 30 min. After rinsing three times with PBS-Tween-20, nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI) and the cells were analyzed by inverted fluorescence microscopy.

Two siRNAs against COTE1 were designed using the web server of Invitrogen Co. and chemically synthesized by Shanghai GenePharma Co. The commonly used negative control (NC) siRNA supplied by Qiagen was used as a control. The sense and antisense sequences of human COTE1 were as follows: siRNA-2852 sequence: 5′-GUAUGUAAGCCUUCAAUAAdTdT-3′ (sense) and 5′-UUA UUGAAGGCUUACAUACdTdT-3′ (antisense); siRNA-3129 sequence: 5′-AGCUCUUAACAGUAUGUAAdTdT-3′ (sense) and 5′-UUACAUACUGUUAAGAGCUdTdT-3′ (antisense).

To construct RNAi plasmid, the shRNA expression cassette containing sequence identical to the siRNA2852/3129 sequence of the target gene was inserted into the expression plasmid pSUPER containing the polymerase-III H1-RNA gene promoter. Negative control oligos served as controls. For the construction of COTE1 recombinant plasmid, the COTE1 open reading frame was amplified from a human liver cDNA library (Genbank: NM_006589.2) using nested PCR and inserted into pcDNA3.1B-FLAG-GFP (Chinese National Human Genome Center, Shanghai). The sequences of the shRNAs and primers used for COTE1 expression vector construction are shown in Table I.

RNAi and plasmid transfections were performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions at a cell density of 30–50% and 80–90%, respectively.

For the cell proliferation assay, cells were plated in 96-well plates and cultured for 24 h to a density of 30–50% before transient transfection. The cell counting kit-8 (CCK-8; Dojindo Labs.) was used to measure cell viability according to the manufacturer’s instructions. Briefly, 90 μl DMEM free medium plus 10 μl CCK-8 solution was added to the 96-well plate. After incubation at 37°C for 1 h, the absorbance at 450 nm was measured. Three replicate wells were tested per assay condition, and each experiment was repeated at least three times.

For the colony formation assay, plasmid-transfected cells were plated in 100 mm plate and cultured for 2–3 weeks until visible colonies were present. Cells were cultured in media containing fetal bovine serum and G418 (Life Technologies, Inc.) at a final concentration of 0.6–1 mg/ml. Colonies were stained with Coomassie Brillant Blue R-250 (CBBR-250) for 1 h. All experiments were repeated independently at least three times.

For the soft agar colony formation assay, 3000 cells were plated in each well of a 24-well plate containing 1% base agar and 0.5% top agar and cultured at 37°C for 2–3 weeks. Colonies were counted under a dissecting microscope. All experiments were repeated independently at least three times.

To establish the tumor xenograft model, we first generated stable transfected cell lines. Briefly, Focus and WRL-68 HCC cells were transfected with pcDNA3.1B-cote1-Flag and pSUPER-shRNA2852, respectively, and grown in the presence of G418 (0.6–1 mg/ml) for 2–3 weeks. Visible colonies were picked and transferred into 96-well plates with DMEM++ medium and G418. Stable transfected cells in the 96-well plate were digested and successively transferred into 24-well plates, 6-well plates, and 100-mm plates. Stable cells were injected subcutaneously into both flanks of nude mice (male BALB/c, 4–6 weeks-old). Tumor volume for each mouse was determined by measuring in two dimensions and calculated as: tumor volume = length × (width)²/2. The tumor tissues were formalin-fixed and paraffin-embedded for immunohistochemistry.

COTE1 expression in tissues of HCC specimens or nude mouse tumor xenografts was determined by immunohistochemistry. Briefly, formalin-fixed samples were paraffin-embedded and cut into 4-μm sections. Slides were incubated with goat anti-COTE1 polyclonal antibody (1:50; Santa Cruz Biotechnology) or normal goat IgG as a negative control at 4°C overnight. For detection, MaxVision™ HRP-Polymer anti-Goat IHC Kit (Maixin Bio. Ltd., China) was used according to the manufacturer’s protocol. Stained slides were observed under a light microscope.

Transfected cells were harvested at different time points, fixed in cold 70% ethanol, washed, rehydrated in PBS, and stained with propidium iodide (PI) binding buffer (10 mg/ml RNase A and 10 μg/ml PI) for 30 min at room temperature. DNA content of the cells was analyzed using a FACSCalibur flow cytometer (Becton-Dickinson) with collection of 10,000 events. Analyses of cell cycle and apoptotic cells (sub-G1 population) were performed using Becton-Dickinson FACScan.

Cells were fixed in 4% methanol-free formaldehyde solution in PBS (pH 7.4) for 25 min at 4°C, washed with PBS for 10 min at room temperature, and permeabilized in 0.2% Triton X-100 solution in PBS for further 5 min. After equilibration for 10 min, the cells were incubated with rTdT (Promega) and observed under a fluorescence microscope. A nucleus with bright green fluorescent staining was recorded as a TUNEL-positive event (11).

Cell lysates were prepared in cold lysis buffer containing 25 mmol/l Tris-Cl (pH 7.5), 5 mmol/l EDTA, 1% sodium dodecyl sulfate, and protease inhibitor cocktail (Sigma). After boiling for 5 min, samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto a polyvinylidene difluoride membrane, which was blocked in 5% blocking buffer for 2 h at room temperature. The membrane was incubated with primary antibody in PBS-Tween-20 (0.1% Tween-20 in PBS) at 4°C overnight and with secondary antibody at room temperature for 1 h. Antibodies used in this study were: goat anti-COTE1 (1:200; Santa Cruz Biotechnology), mouse anti-WWOX (1:200; Santa Cruz Biotechnology), rabbit anti-WWOX (phospho Y33, phospho Y287, 1:500; Abcam, UK), mouse anti-flag (1:200; Santa Cruz Biotechnology), rabbit anti-cyclin D1, E1 (1:500; Abcam), rabbit anti-p53 (p53 and phospho S46, 1:500; Abcam), rabbit anti-Bcl-2 (1:500; Abcam), rabbit anti-caspase 3, 9 (1:500; Abcam), and anti-β-actin (1:500; Santa Cruz Biotechnology). Proteins were detected using the Odyssey Infared Imaging System (Li-COR).

Cells transfected with pcDNA3.1-COTE1-Flag were resuspended in 1 ml lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1.0% Triton X-100, 1 mM EDTA, and protease inhibitor cocktail). Immunoprecipitation of lysates was conducted using anti-Flag antibody (1:100; Santa Cruz Biotechnology), followed by immunoblotting with antibodies against WWOX (1:200; Santa Cruz Biotechnology) or COTE1 (1:100; Santa Cruz Biotechnology). Lysate of cells transfected with empty vector (pcDNA3.1) served as a control (12).

All quantitative data were recorded as means ± standard deviation (SD). Differences between two groups were assessed by Student’s t-test (two-tailed) using GraphPad PRISM 5 software. Comparisons among multiple groups were performed by one-way analysis of variance and least significant difference t-test. Categorical data were evaluated by the χ² test. In all tests, p<0.05 was considered statistically significant.

Previous gene microarray analysis performed in our laboratory showed that COTE1 was upregulated in 11 paired HCC tissues (data not shown). To confirm these findings, we performed RT-PCR and quantitative PCR to measure the mRNA expression level of COTE1 in 80 paired HCC clinical specimens relative to the levels in corresponding adjacent non-cancerous liver. The results clearly showed upregulation of COTE1 mRNA in HCC specimens (Fig. 1A). To evaluate the protein level of COTE1 in HCC, we performed immunochemical staining with a specific antibody against COTE1 in another 80 matched samples. In non-HCC tissues 24/80 (30%) showed no or weak (+/−) positive staining with the rest showed no staining, whereas 37/80 (46.25%) of cancer specimens showed positive staining: seven mildly positive, 17 moderately positive, and 13 strongly positive (Fig. 1B). We also evaluated the expression pattern of COTE1 in HCC cell lines by RT-PCR and found that COTE1 was highly expressed in 11 of 15 HCC cells relative to expression in normal human adult liver tissue (Fig. 1C). Together, these data showed that COTE1 is upregulated in HCC, confirming the results of the gene microarray analysis.

To investigate the relationship between COTE1 expression and HCC clinical features, we further analyzed the results of RT-PCR in the 80 HCC specimens. The clinical characteristics of the patients and tumors are summarized in Table II. The prepared specimens were grouped by gender (male or female), age (≥45 years or <45 years), etiology (HBV⁺ or HBV⁻), tumor size (≥3 cm or <3 cm), metastasis (yes or no) and Edmondson grade (I–II or III–IV). The resulting data showed that increased transcriptional expression of COTE1 was statistically correlated with the pathological tumor size and Edmondson grade (p<0.05). However, no statistical correlation was found for the other variables. Taken together, these data indicate that upregulation of COTE1 contributes to HCC growth and poor differentiation.

To identify the subcellular localization of COTE1 in HCC cells we performed immunofluorescence staining of endogenous COTE1 in PLC/PRF/5 cells. The results showed localization of COTE1 predominantly in the cytoplasm and to a lesser extent in the cell membrane of HCC cells (Fig. 1D).

To investigate whether COTE1 contributes to hepatocarcinogenesis, we investigated the effect of COTE1 on cell proliferation and colony formation. Based on the expression pattern of COTE1 in HCC cell lines, recombinant pcDNA3.1-COTE1-Flag was transiently transfected into Focus, Huh7, MHCC-LM6, and MHCC-L cells, all of which express a relatively low level of COTE1 (Fig. 1C). The empty vector pcDNA3.1-Flag was used as a control. Interestingly, cellular growth of Focus and Huh7 cells was significantly promoted by exogenous COTE1 compared with cells transfected with empty vector (Fig. 2A and B) but a similar response was not observed in MHCC-LM6 and MHCC-L. To further investigate the long-term effect of COTE1 on cell proliferation, transfected Focus and Huh7 cells were cultured in G418 for 2–3 weeks and colony formation was measured. Few colonies formed for cells transfected with vector alone, whereas colonies were more visible and a greater number of colonies formed in COTE1-overexpressing cells (Fig. 2C and D). These data suggest that ectopic COTE1 expression selectively enhanced the viability of HCC cells.

We further evaluated the effect of COTE1 on cellular proliferation using YY-8103, WRL-68, PLC, and MHCC-97H cells, which showed high COTE1 expression levels (Fig. 1C). These cells were transfected with two chemically synthesized siRNAs that target COTE1, siRNA-2852 and siRNA-3129. As expected, endogenous COTE1 expression was efficiently knocked down in all cell lines. Proliferation was significantly inhibited in YY-8103 and WRL-68 cells transfected with siRNAs, but not in PLC or MHCC-97H cells (Fig. 3A-C). To further test the effect of COTE1 on cellular growth, we performed a soft agar colony formation assay, which more closely imitates in vivo growth. Given the short half-life of siRNAs, we used shRNAs derived from recombinant pSUPER vector for these experiments. YY-8103 and WRL-68 cells were transfected with shRNA and cultured in soft agar for 2–3 weeks. As expected, cells transfected with shRNA-2852/3129 produced fewer colonies than those transfected with shRNA-NC plasmid (Fig. 3D-F). These collective data indicate that endogenous COTE1 plays an essential role in cellular proliferation and colony formation of HCC cells.

To assess the effect of COTE1 on HCC cell proliferation in vivo, stable COTE1-transfected Focus or WRL-68 cells were implanted into one flank of nude mice and negative control cells were injected into the other flank. As expected, transfection with pcDNA3.1B-COTE1-Flag markedly enhanced tumorigenicity of cells compared with the controls (Fig. 4A and D, upper), whereas shRNA2852 suppressed tumorigenicity (Fig. 4B and D, lower). Immunohistochemical analysis confirmed elevated COTE1 expression in the tumors formed by COTE1-transfected cells, and reduced COTE1 expression in tumors formed by cells transfected with COTE1 shRNA (Fig. 4C).

We first performed flow cytometry to evaluate cell cycle distribution of Focus, Huh7, YY-8103, and WRL-68 cells 24 and 48 h after transfection with siRNA2852. G0/G1 phase arrest was obvious in YY-8103 cells 48 h after transfection (Fig. 5A and B). In contrast, G1- to S-phase transition was evident in Huh7 cells at 24 h and reached a peak at 48 h post-transfection (Fig. 5C and D).

We next performed flow cytometry and TUNEL assays to determine the effect of COTE1 on apoptosis under reduced serum conditions. First, we investigated the apoptosis of YY-8103 and WRL-68 cells at different time points after shRNA transfection. Compared with the control group, COTE1 inhibition resulted in a larger percentage of sub-G1 cells in WRL-68, but not YY-8103 cells, reaching a peak at 72 h (Fig. 6A). To obtain further evidence of apoptosis, we performed the TUNEL assay in WRL-68 cells at 72 h post-transfection and obtained results consistent with those of flow cytometry (Fig. 6B and C). The effect of COTE1 upregulation on apoptosis was examined by treatment of COTE1-overexpressing Focus and Huh7 cells with doxorubicin. As shown in Fig. 6D and E, doxorubicin (5 μM, 24 h) treatment resulted in a larger proportion of apoptotic cells in Focus cells transfected with empty vector compared with the COTE1 overexpressing cells.

The data presented above suggest that COTE1 functions as a novel oncogene in HCC, since COTE1 inhibition obviously suppresses cell proliferation via cell cycle arrest and apoptosis. Considering previous reports that COTE1 binds to the WW domain of WWOX in vitro (7,9), we investigated whether the effect of COTE1 on cell proliferation was mediated by WWOX. First, we carried out co-localization experiments to determine whether COTE1 might interact with WWOX, and observed COTE1-WWOX co-localization in the cytoplasm of Focus and Huh7 cells by fluorescence microscopy (Fig. 7A). Next, we performed co-immunoprecipitation (co-IP) assays to determine whether COTE1 and WWOX physically interact in Focus and Huh7 cells transfected with pcDNA3.1-COTE1-Flag. The mutual co-IP data indicated that COTE1 physically associates with WWOX (Fig. 7B). To explore the effect of COTE1 on WWOX, we measured tyrosine phosphorylation of WWOX (Tyr33 and Tyr287) in transfected Focus, Huh7, YY-8103, and WRL-68 cells. Our data showed that the level of pTyr33 was decreased in Focus cells and increased in WRL-68; whereas the level of pTyr287 was reduced in Huh7 cells and increased in YY-8103 (Fig. 7C). We also measured p-p53 (Ser46), which participates in WWOX-mediated apoptosis, and found that p-p53 levels were suppressed in Focus cells, and increased in WRL-68 (Fig. 7C).

To validate the hypothesis that COTE1-pWWOX mediates mitochondrial apoptosis in HCC cells, we performed western blot analysis to measure the expression of molecules involved in the mitochondrial apoptosis pathway: Bax, Bcl-2, caspase-9, and caspase-3. Our data showed that upregulation of pWWOX (Tyr33) by COTE1 knockdown induced p53 (Ser46)-mediated endogenous apoptosis of WRL-68 cells and, conversely, Tyr33 dephosphorylation of WWOX by COTE1 overexpression rendered Focus cells resistant to p53 (Ser46)-mediated apoptosis (Fig. 8A).

We then measured expression of cyclin D1 and cyclin E1, both of which function as positive regulators in promoting G1- to S-phase transition, in Huh7 and YY-8103 cells and found that expression of cyclin D1, but not cyclin E1, negatively correlated with COTE1 expression. However, there was a positive correlation between cyclin D1 and pWWOX (Tyr287) (Fig. 8B). These data suggest that COTE1-pWWOX (Tyr287)-cyclin D1 mediate cell cycle regulation in HCC.

In conclusion, the above findings indicate that COTE1 regulates the WWOX-p53-mediated apoptosis pathway through Tyr33 dephosphorylation and participates in WWOX-induced cell cycle progression via Tyr287 dephosphorylation (Fig. 8C).