Translationally controlled tumor protein induces mitotic defects and chromosome missegregation in hepatocellular carcinoma development

Authors

  • Tim Hon Man Chan,

    1. State Key Laboratory for Liver ResearchThe University of Hong Kong, Pokfulam, Hong Kong, China
    2. Departments of Clinical OncologyThe University of Hong Kong, Pokfulam, Hong Kong, China
    Search for more papers by this author
  • Leilei Chen,

    1. State Key Laboratory for Liver ResearchThe University of Hong Kong, Pokfulam, Hong Kong, China
    2. Departments of Clinical OncologyThe University of Hong Kong, Pokfulam, Hong Kong, China
    Search for more papers by this author
  • Ming Liu,

    1. State Key Laboratory for Liver ResearchThe University of Hong Kong, Pokfulam, Hong Kong, China
    Search for more papers by this author
  • Liang Hu,

    1. State Key Laboratory for Liver ResearchThe University of Hong Kong, Pokfulam, Hong Kong, China
    Search for more papers by this author
  • Bo-jian Zheng,

    1. MicrobiologyThe University of Hong Kong, Pokfulam, Hong Kong, China
    Search for more papers by this author
  • Vincent Kwok-Man Poon,

    1. MicrobiologyThe University of Hong Kong, Pokfulam, Hong Kong, China
    Search for more papers by this author
  • Pinzhu Huang,

    1. State Key Laboratory of Oncology in Southern China, Sun Yat-Sen University Cancer Center, Guangzhou, China
    Search for more papers by this author
  • Yun-Fei Yuan,

    1. State Key Laboratory of Oncology in Southern China, Sun Yat-Sen University Cancer Center, Guangzhou, China
    Search for more papers by this author
  • Jian-dong Huang,

    1. BiochemistryThe University of Hong Kong, Pokfulam, Hong Kong, China
    Search for more papers by this author
  • Jie Yang,

    1. Anatomy, The University of Hong Kong, Pokfulam, Hong Kong, China
    Search for more papers by this author
  • George Sai-wah Tsao,

    1. Anatomy, The University of Hong Kong, Pokfulam, Hong Kong, China
    Search for more papers by this author
  • Xin-Yuan Guan

    Corresponding author
    1. State Key Laboratory for Liver ResearchThe University of Hong Kong, Pokfulam, Hong Kong, China
    2. Departments of Clinical OncologyThe University of Hong Kong, Pokfulam, Hong Kong, China
    3. State Key Laboratory of Oncology in Southern China, Sun Yat-Sen University Cancer Center, Guangzhou, China
    4. BiochemistryThe University of Hong Kong, Pokfulam, Hong Kong, China
    • State Key Laboratory for Liver Research, Departments of Clinical Oncology, Biochemistry, and Anatomy, The University of Hong Kong, Room L10-56, Laboratory Block, 21 Sassoon Road, Pokfulam, Hong Kong, China
    Search for more papers by this author
    • fax: (852) 2872 5197


  • Potential conflict of interest: Nothing to report.

  • This work was supported by a Hong Kong Research Grant Council Grant (HKU 7656/07M), a Hong Kong Research Grant Council Central Allocation (HKU5/CRF/08), a Hong Kong RGC Collaborative Research Grant (HKU 7/CRG09), the “Hundred Talents Program” at Sun Yat-Sen University (85000-3171311), the Major State Basic Research Program of China (2006CB910104), and grants from the National Natural Science Foundation of China (30772475 and 30971606).

Abstract

Emerging evidence implicates the chromodomain helicase/ATPase DNA binding protein 1–like gene (CHD1L) as a specific oncogene in human hepatocellular carcinoma (HCC). To better understand the molecular mechanisms underlying HCC cases carrying CHD1L amplification (>50% HCCs), we identified a CHD1L target, translationally controlled tumor protein (TCTP), and investigated its role in HCC progression. Here, we report that CHD1L protein directly binds to the promoter region (nt −733 to −1,027) of TCTP and activates TCTP transcription. Overexpression of TCTP was detected in 40.7% of human HCC samples analyzed and positively correlated with CHD1L overexpression. Clinically, overexpression of TCTP was significantly associated with the advanced tumor stage (P = 0.037) and overall survival time of HCC patients (P = 0.034). In multivariate analyses, TCTP was determined to be an independent marker associated with poor prognostic outcomes. In vitro and in vivo functional studies in mice showed that TCTP has tumorigenic abilities, and overexpression of TCTP induced by CHD1L contributed to the mitotic defects of tumor cells. Further mechanistic studies demonstrated that TCTP promoted the ubiquitin-proteasome degradation of Cdc25C during mitotic progression, which caused the failure in the dephosphorylation of Cdk1 on Tyr15 and decreased Cdk1 activity. As a consequence, the sudden drop of Cdk1 activity in mitosis induced a faster mitotic exit and chromosome missegregation, which led to chromosomal instability. The depletion experiment proved that the tumorigenicity of TCTP was linked to its role in mitotic defects. Conclusion: Collectively, we reveal a novel molecular pathway (CHD1L/TCTP/Cdc25C/Cdk1), which causes the malignant transformation of hepatocytes with the phenotypes of accelerated mitotic progression and the production of aneuploidy. (HEPATOLOGY 2012)

Hepatocellular carcinoma (HCC) is the sixth most common human cancer in the world, with extremely poor prognosis and a <3% 5-year survival rate for untreated cancer.1 The ultimate cause of HCC is perhaps better understood than other types of human cancers, which is chronic liver disease (eventually leading to cirrhosis), particularly chronic hepatitis B and C and alcoholic liver disease. Other risk factors, such as tobacco smoking, nonalcoholic steatohepatitis, and inherited metabolic diseases, have also been proposed to cause HCC, albeit at a lower frequency.2 In addition, HCC is predominantly male associated in all populations, and the incidence of HCC also increases progressively with age. Noninvasive, low-cost diagnostic tools serum alpha-fetoprotein (AFP) and ultrasonography (US) are available for a well-defined cirrhotic population and satisfy almost all requisites, but their diagnostic accuracy has been largely debated.2, 3 Patients with preserved liver function and either a solitary nodule <5 cm or up to three nodules <3 cm each are eligible for curative treatments, including surgical resection, liver transplantation (LT), and percutaneous ablation. Percutaneous ethanol injection (PEI), radiofrequency ablation (RFA), or transarterial chemoembolization (TACE) are safe and effective in bridging patients with HCC to LT.

During HCC progression, amplification of chromosome 1q21 has been detected in 58%-78% primary HCC cases,4 suggesting that one or more oncogenes within the amplicon play critical role in HCC development. Recently, we isolated a candidate oncogene, chromodomain helicase/ATPase DNA binding protein 1–like gene (CHD1L; previously called ALC1), within the 1q21 region by hybrid selection using microdissected DNA from this region.5 Previous in vivo and in vitro studies demonstrate that CHD1L is a critical HCC-associated oncogene and induces cellular malignant transformations.5, 6 In addition, spontaneous tumor formation has been found in 10 of 41 of CHD1L-transgenic mice, including 4 mice with HCCs, but not in their 39 wild-type littermates.7 To explore the regulatory network in which CHD1L contributes to HCC development, CHD1L-regulated proteome was characterized by two-dimensional electrophoresis (2DE) and mass spectrometry (MS). One up-regulated protein, TCTP, was selected for further characterization. TCTP is a housekeeping gene expressed in almost all mammalian tissues and is highly conserved among animals, plants, and yeast. Based on its amino acid sequence, TCTP, also named p23, cannot be attributed to any known protein family. TCTP was first found in Ehrlich ascites tumor cells, and overexpression of TCTP has been detected in liver and colorectal cancers.8, 9 The first overexpression experiment in the analysis of TCTP showed that it interacts with microtubules in a cell-cycle–dependent manner.10 It has also been reported that TCTP functions as a prosurvival factor by promoting cell cycle and inhibiting apoptosis.10, 11 However, the underlying mechanism of TCTP overexpression in cancers and the precise mechanism by which TCTP regulates cell-cycle progression are far from clear.

The aim of this study was to assess the clinical significance of TCTP in human HCC and to identify mechanisms mediating the overexpression of TCTP, with a focus on its cell-cycle–modulatory function, and to reveal a molecular mechanism linking increased TCTP expression to cancer progression.

Abbreviations

AFP, alpha-fetoprotein; bp, base pairs; ChIP, chromatin immunoprecipitation; CHD1L, chromodomain helicase/ATPase DNA binding protein 1–like gene; CHX, cycloheximide; CIN, chromosome instability; CI, confidence interval; DIG, digoxigenin; GFP, green fluorescent protein; HR, hazard ratio; HA-Ub, hemagglutinin-tagged ubiquitin; HCC, hepatocellular carcinoma; Ig, immunoglobulin G; IHC, immunohistochemical; LT, liver transplantation; MS, mass spectrometry; OS, overall survival; PEI, percutaneous ethanol injection; qPCR, quantitative PCR; pRL-TK, Renilla luciferase plasmid; RFA, radiofrequency ablation; SD, standard deviation; shRNA, short hairpin RNA; siRNA, short interfering RNA; TACE, transarterial chemoembolization; TCTP, translationally controlled tumor protein; 2DE, two-dimensional electrophoresis; US, ultrasonography.

Patients and Methods

Patients and Clinical Specimens.

HCC specimens were collected immediately after hepatectomy at the Sun Yat-Sen University Cancer Center (Guangzhou, China) from 2001 to 2010. None of these patients received preoperative chemotherapy or radiotherapy. All HCC patients gave written informed consents on the use of clinical specimens for medical research. Samples used in this study were approved by the Committees for Ethical Review of Research at Sun Yat-Sen University.

Cell Lines.

HCC cell lines QGY-7703, BEL7402, PLC8024, Hep3B, Huh7, HepG2, and an immortalized normal human liver cell line (LO-2) was obtained from the Institute of Virology, Chinese Academy of Medical Sciences (Beijing, China).

Chromatin Immunoprecipitation/Polymerase Chain Reaction.

Chromatin immunoprecipitation (ChIP) experiments were performed using an EZ-Magna ChIP G kit (Upstate Biotechnology, Lake Placid, NY), according to the manufacturer's instructions. CHD1L-binding DNA fragments were pulled down by two different anti-GFP (green fluorescent protein) antibodies (FL and B-2) or pooled immunoglobulin G (IgG) from mouse and rabbit (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) as a negative control.

Supershift Assay.

Nuclear extract was prepared using the NucBuster Protein Extraction kit (Novagen Inc., San Diego, CA). Probes were end-labeled with digoxigenin (DIG) by polymerase chain reaction (PCR), using DIG-labeled deoxyuridine triphosphate (Roche, Indianapolis, IN), in addition to deoxynucleotide triphosphates, then purified by a PCR Purification kit (QIAGEN Inc., Valencia, CA). A supershift assay was performed with 10 μg of nuclear extracts and 50 ng of DIG-labeled or unlabeled probes in 1× binding buffer provided by the Bandshift kit (Amersham Pharmacia Biotech Inc., Piscataway, NJ) and mouse anti-GFP antibody (B-2) (Santa Cruz).

Dual-Luciferase Reporter Assay.

Luciferase reporter constructs (10:10:1 mixture of TCTP luciferase constructs, pcDNA3.1-CHD1L, and Renilla luciferase plasmid; pRL-TK) were tranfected into cells using Lipofectamine (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. Dual-luciferase assays (Promega, Madison, WI) were used to measure luciferase activities, according to the manufacturer's instructions. Results were normalized to the pGL3-basic activity. Sequences of primers used for luciferase reporter constructs are listed in Supporting Table 2.

Cell-Cycle Analysis by Flow Cytometry.

Cell-cycle distribution was examined by flow cytometry using a FACScan flow cytometer (BD Biosciences, Franklin Lakes, NJ). The relative portion of cells in each phase of cell cycle was analyzed using the Modfit program (Verity Software House, Inc., Topsham, ME).

Foci Formation Assay.

For foci formation assay, 1 × 103 of cells were seeded in a six-well plate. After culture for 7 days, surviving colonies (>50 cells/colony) were counted with Gimesa (Invitrogen) staining. Triplicate independent experiments were performed.

In Vivo Ubiquitination Assay.

Cells were transfected with plasmids expressing Cdc25C (OriGene, Rockville, MD), TCTP, and hemagglutinin (HA)-tagged ubiquitin (HA-Ub) (Sigma-Aldrich, St. Louis, MO), either alone or in combination. For inhibition of proteasome-mediated protein degradation, cells were treated with 20 μM proteasome inhibitor MG132 (Calbiochem, San Diego, CA) for 6 hours before harvest. Cell lysates were analyzed by western blotting with anti-Cdc25C, anti-TCTP, or anti-HA (Santa Cruz) antibodies. Remaining cell lysates were immunoprecipitated with 5 μg of anti-Cdc25C antibody and 100 μL of protein G agarose provided by the immunoprecipitation kit (Roche). For analysis of Cdc25C stability in metaphase, cells were synchronized at prometaphase and then released in completed medium with 50 μg/mL of cycloheximide (CHX; Calbiochem) and harvested at 0, 30, 60, 120, and 300 minutes.

In Vivo Tumorigenicity Assay.

Approximately 2 × 106 of TCTP-7703 or Vec-7703 cells were injected subcutaneously into the right or left dorsal flank of 4-5-week-old BALB/cAnN-nu (nude) mice, respectively. Tumor volume was measured weekly and calculated by the following formula: V = 0.5 × W2 × L. All animal experiments were approved by, and performed in accord with, the Committee of the Use of Live Animals in Teaching and Research at the University of Hong Kong (Pokfulam, Hong Kong, China).

Xenograft Tumor Processing.

Xenograft tumor samples were thoroughly washed and minced into ∼1 mm3 pieces and incubated in 1× Accumax (Innovative Cell Technologies, Inc., San Diego, CA) diluted in phosphate-buffered saline, according to the manufacturer's instructions. Single-cell suspension was obtained by filtering the supernatant through 100 μm, followed by a 40-μm cell strainer (BD).

Time-Lapse Microscopy.

Approximately 30%-40% confluent cells were seeded in 35-mm diameter CELLview dishes (Greiner Bio-One GmbH, Frickenhausen, Germany). Cells were observed using the PerkinElmer Spinning Confocal/Widefield Imaging system (PerkinElmer, Waltham, MA). Time-lapse images were recorded at 3-minute intervals for 24 hours with a 63× objective lens. Image analysis was performed using Metamorph off-line software and ImageJ.

Statistical Analysis.

Clinicopathological features in patients with overexpression and patients without overexpression were compared using nonparametric cross-tabs analysis (chi-square test or Fisher's exact test) for categorical variables. Based on the fold-change values of TCTP, TCTP expression levels in HCC tissues and their matched nontumor tissues were compared using the Wilcoxon signed-rank test. Kaplan-Meier plots and log-rank tests were used for survival analysis. Spearman correlation coefficients were used to evaluate the positive correlation between CHD1L and TCTP in clinical samples. The independent Student's t test was used to compare number of foci, tumor size, and luciferase activity between any two preselected groups. A P value less than 0.05 was considered statistically significant.

Results

Identification of CHD1L-Regulated Proteins by 2DE.

By 2DE and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry mass spectral measurement, a total of eight proteins were differentially expressed in CHD1L-transfected QGY-7703 cells (CHD1L-7703), when compared to empty vector-transfected cells (Vec-7703) (Supporting Fig. 1A; Supporting Table 1). Further validations suggested that TCTP might be a target gene of CHD1L (Supporting Fig. 1B,C).

CHD1L Transcriptionally Up-regulates TCTP Expression.

As reported by our previous study, CHD1L possesses a DNA-binding activity with a putative DNA-binding motif (C/A)C(T/A)T(T/A/G)T,12 and CHD1L may, therefore, up-regulate TCTP expression through a protein-DNA interaction. Using the MatInspector Professional software (Genomatrix Software GmbH, Munich, Germany)13, 14 to search for a CHD1L-binding site within a 1.8-kb upstream region of TCTP, two CHD1L-binding sites were identified at −748 bp (base pairs) and −851 bp in the 5′-flanking region of TCTP (Fig. 1A). By using the ChIP-PCR assay, we found that only DNA Fragment C (nt −733/−1027) containing two CHD1L-binding motifs, but not Fragment A (nt +91/−213) and Fragment B (nt −195/−500), could be detected in CHD1L-ChIPed DNA fragments (Fig. 1B). Supershift signal was only observed in the lane containing DIG-labeled Fragment C, nuclear extract of GFP/CHD1L-7703-C3 cells, and anti-GFP antibody (lane 11) (Fig. 1C).

Figure 1.

CHD1L activates TCTP transcription by binding to the 5′-upstream region (−733/−1027) of TCTP. (A) Schematic diagram representing the distribution of CHD1L-binding loci in the regulatory region of TCTP. Four individual fragments representing 5′-upstream sequences of TCTP (Fragment A: nt −213/+91; Fragment B: nt −500/−195; Fragment C: nt −1027/−733; Fragment D: nt −1027/+91). (B) ChIP assay was performed in an ectopic CHD1L-expressing clone (GFP/CHD1L-7703-C3), as described previously.10 PCR was used to detect Fragments A, B, and C in DNA fragments pulled down by GFP antibodies (FL and B-2) against GFP/CHD1L fusion protein. IgG as a negative control. (C) Supershift assay was used to detect interaction between CHD1L and TCTP double-stranded DNA probes. Lane 1, 5, or 9: mixture of anti-GFP antibody and DIG-labeled probe; lane 2, 6, or 10: mixture of nuclear extract (NE) from GFP-7703-C4 cells, DIG-labeled probe, and anti-GFP antibody; lane 3, 7, or 11: mixture of NE from GFP/CHD1L-7703-C3 cells, DIG-labeled probe, and anti-GFP antibody; lane 4, 8, or 12: lane 3, 7, or 11, plus a 50-fold molar excess of unlabeled probe. (D) Each of the fragments was subcloned into pGL3-basic vector and cotransfected with pcDNA3.1-CHD1L or empty vector into QGY-7703 cells for luciferase reporter assay. All experiments were done in triplicate and repeated at least three times. Data are presented as mean ± standard deviation (SD) (*P < 0.05; **P < 0.001, independent Student's t test). (E) Expression levels of CHD1L and TCTP were determined in scrambled siRNA- or siCHD1L-treated Huh7 and PLC8024 cells by qPCR, and the fold change of expression was calculated from siCHD1L relative to the scrambled. Bars represent the mean ± SD of three independent experiments. (F) Correlation between CHD1L and TCTP expression detected by qPCR in 43 HCC tissues and their paired nontumor tissues, with a linear regression line and Spearmen correlation significance (P < 0.0001).

Luciferase reporter assay was used to further confirm that CHD1L could activate TCTP transcription. As a result, luciferase activity of pGL3-TCTP-FD was significantly increased in cells cotransfected with pcDNA3.1-CHD1L, but not with pcDNA3.1 (P < 0.05), whereas the luciferase activities of pGL3-TCTP-FA, FB, and FC were not increased in cells cotransfected with pcDNA3.1-CHD1L (Fig. 1D). These results demonstrated that CHD1L could bind to the CHD1L-binding motifs within the 5′-upstream region (nt −733/−1027) of TCTP and activate TCTP transcription; however, the other sequence (Fragment A: nt +91/−213) of the 5′-upstream region was also required for the activation of TCTP transcription.

Expression of TCTP Is Positively Correlated With CHD1L Expression.

Protein expression of TCTP and CHD1L was detected in seven cell lines, including six HCC cell lines and one immortalized human liver cell line (LO-2). TCTP expression was positively correlated with that of CHD1L in these seven cell lines (Spearmen correlation coefficient, 0.786; P = 0.048) (Supporting Fig. 2A-C). Serial sections of 5 HCCs with surrounding nontumor tissues were stained with antibodies against CHD1L and TCTP. As a result, the expression patterns of TCTP and CHD1L showed almost perfect concordance in both tumor and nontumor tissues (Supporting Fig. 2D). Furthermore, PLC8024 and Huh7 cells were treated with short interfering RNA (siRNA) against CHD1L (siCHD1L) or the corresponding scrambled siRNA. As detected by quantitative PCR (qPCR), siCHD1L-treated cells showed the lower expression of both CHD1L and TCTP than that of the scrambled siRNA-treated cells (P < 0.0001 and P < 0.00001; Fig. 1E), suggesting that TCTP expression was modulated in a CHD1L-dependent manner. We further investigated the positive correlation between CHD1L and TCTP in clinical specimens. TCTP expression was significantly correlated with CHD1L expression in these specimens (Spearmen correlation coefficient, 0.449; P < 0.0001; Fig. 1F), further indicating that CHD1L is able to up-regulate TCTP expression.

Clinical Significance of TCTP in HCC.

To determine the prevalence and clinical significance of TCTP in HCC, the correlation between overexpression of TCTP and the clinicopathological features was investigated in a retrospective cohort of 118 HCC patients. As detected by qPCR, overexpression of TCTP (defined as a greater than 2-fold increase) was detected in 40.7% (48 of 118) of HCC cases. HCC tissues showed higher expression of TCTP than adjacent nontumor tissues (Wilcoxon signed rank test, P = 0.0336; Fig. 2A). Overexpression of TCTP in HCC tissues was significantly associated with advanced tumor stage (P = 0.037; Table 1). To confirm our findings, immunohistochemical (IHC) staining of TCTP was conducted in paraffin sections from 20 patients with HCCs of different tumor stages (6 HCCs of stage I, 6 HCCs of stage II, and 8 HCCs of stage III). In 9 of 14 (57.1%) of advanced HCC cases (stage II and III), expression of TCTP was obviously higher in tumor tissues, as compared to their adjacent nontumor tissues (Fig. 2C), whereas 5 of 6 (83.3%) of stage I HCC tissues showed an expression pattern of TCTP similar to nontumor tissues (Fig. 2B). The prognostic significance of TCTP overexpression was also studied in this cohort of 108 patients with valid follow-up data. As a result, TCTP overexpression was significantly associated with shorter overall survival (OS) of patients (log rank = 4.495, P = 0.034; Fig. 2D). In univariate analyses, statistically significant predictors for patient survival were vascular invasion, cell differentiation status, American Joint Committee on Cancer tumor staging, and TCTP expression level (Fig. 2E). In multivariate analyses, TCTP expression level demonstrated better predictive power for patient survival (hazard ratio [HR]: 2.488; 95% CI: 1.020-6.068; P = 0.048, Fig. 2E) than other predictors.

Figure 2.

Clinical significance of TCTP in human HCC. (A) Scatter plots of fold change of TCTP in HCC and their matched nontumor tissues (P = 0.0336; Wilcoxon signed-rank test). In both panels, black lines indicate mean ± SD. (B and C) Representative images of hematoxylin and eosin staining and IHC analysis of TCTP expression on serial sections of HCC tissues (B, stage I; C, stage III) (upper panels; 100× magnification). Tumor (T) and adjacent nontumor (NT) regions are magnified and shown in lower panels (200× magnification). (D) Kaplan-Meier OS curve of HCC patients in correlation with overexpression (OE) of TCTP. (E) Multivariate Cox regression analysis of OS for TCTP OE in the cohort. HRs, CI, and P values are shown for all characteristics, as listed.

Table 1. Correlation of TCTP Overexpression With Clinicopathological Features In 118 HCCs
Clinical FeaturesNumberTCTP Overexpression (%)P Value
  • Abbreviations: TCTP, translationally controlled tumor protein; HCCs, hepatocellular carcinomas; HBsAG, hepatitis B surface antigen; AFP, alpha-fetoprotein; AJCC, American Joint Committee on Cancer.

  • *

    Partial data were not available, and statistics were based on available data.

  • †Statistical significance (P < 0.05) is shown in bold.

Gender   
 Female209 (45.0) 
 Male9839 (39.8)0.803
Age, years   
 ≦609337 (39.8) 
 >602512 (48.0)0.712
HBsAg*   
 Negative2812 (42.9) 
 Positive8835 (39.7)0.827
    
Serum AFP (ng/mL)*a   
 <=4006628 (42.4) 
 >4004919 (38.8)0.706
Tumor size (cm)   
 ≦53616 (44.4) 
 >58232 (39.0)0.685
    
Cell differentiation*   
 Well differentiated (I-II)3913 (33.3) 
 Moderately differentiated (III)3113 (41.9) 
 Poorly differentiated (IV)31 (33.3)0.753
    
Vascular invasion*   
 Absent8934 (38.2) 
 Present209 (45.0)0.618
Tumor stage (AJCC)*a   
 Stage I5315 (28.3) 
 Stage II3017 (56.7) 
 Stage III3314 (42.4)0.037b

TCTP Has Strong Tumorigenic Ability.

Compared to empty vector-transfected QGY-7703 cells (Vec-7703), two TCTP transfectants (TCTP-C2 and TCTP-C7) showed higher expression levels of TCTP (Supporting Fig. 3A). As expected, TCTP-C2 and C7 cells showed higher frequencies of foci formation, when compared to Vec-7703 cells (P < 0.001; Supporting Fig. 3A; Fig. 3B). Vec-7703 and TCTP-7703 cells (the pool of TCTP-C2 and TCTP-C7) were subcutaneously injected into the left and right dorsal flank of each mouse (n = 6), respectively. Tumor formation was observed in 5 of 6 and 1 of 6 of TCTP-7703 and Vec-7703-injected nude mice, respectively (Fig. 3B). The average volume of tumors induced by TCTP-7703 was significantly larger than that induced by Vec-7703 cells (Fig. 3C).

Figure 3.

TCTP has tumorigenic abilities and decreases cell accumulation in G2/M phase. (A) Numbers of foci induced by TCTP-C2, C7, and Vec-7703 cells were calculated and summarized in the bar chart. Three independent experiments were performed, and data are presented as mean ± SD (**P < 0.001, independent Student's t test). (B) Representatives of tumor (indicated by arrows) formation in nude mice (n = 6) induced by Vec-7703 cells (left dorsal flank) and TCTP-7703 (right dorsal flank). Tumor growth curves in nude mice induced by TCTP-7703 and Vec-7703 cells are summarized in (C) (**P < 0.001; independent Student's t test). (D) Vec-7703 and TCTP-7703 cells were synchronized at early S phase (0 hours) by double thymidine block and released by adding complete medium. DNA content was detected by flow cytometry at the indicated time points. Percentages of cells at G0/G1, S, and G2/M are summarized in (E). The value is expressed as the average of three independent experiments. (F) Western blotting analysis of cyclin B1 expression in Vec-7703 and TCTP-7703 cells at each time point. β-actin was used as a loading control. Quantification of protein expression was obtained by measuring the intensity of the bands with ImageJ software. The value is expressed as the average of two independent experiments.

TCTP Decreases Cell Accumulation in G2/M Phase.

To investigate the mechanism underlying the tumorigenic ability of TCTP, the role of TCTP in cell-cycle progression was studied. As a result, TCTP-7703 and Vec-7703 cells progressed similarly from G1- to S-phase transition (Supporting Fig. 4A,B), and similar temporal expressions of G1/S checkpoints between TCTP-7703 and Vec-7703 cells were detected by western blotting analysis (Supporting Fig. 4C). To study the effect of TCTP on S/G2 transition, cells were synchronized at early S phase by double thymidine block, then released in complete medium. Interestingly, TCTP-7703 cells showed a decreased accumulation in the G2/M phase, compared to Vec-7703 cells. After release for 10 hours post-thymidine block, the percentage of Vec-7703 or TCTP-7703 cells at G2/M phase was 67.53% or 41.90%, respectively (Fig. 3D,E). In addition, the expression level of cyclin B1 in Vec-7703 cells increased gradually and peaked at 12 hours, which, in TCTP-7703 cells, was expressed in lower levels, compared to Vec-7703 cells, and peaked at 10 hours and decreased promptly 12 hours after being released (Fig. 3F). The expression of cyclin B1, known to begin accumulating during late S and G2 phases, peaks at late G2/M phase, starts to degrade at the start of metaphase, and is nearly completely degraded at the onset of anaphase.15 These data suggest that overexpression of TCTP may lead to a faster exit from mitosis.

TCTP Overexpression Promotes a Faster M-Phase Exit and Causes Mitotic Defects and Chromosome Missegregation.

Vec-7703 and TCTP-7703 cells were arrested at prometaphase by thymidine-nocodazole block. After being released from thymidine-nocodazole block, the M-phase exit in TCTP-7703 cells was faster than that of Vec-7703 cells (Fig. 4A,B). To confirm the effect of TCTP on mitotic exit, cells were stained with anti-α-tubulin antibody 1.5 hours after release. As a result, the midbody could be frequently observed in Vec-7703 cells, but not in TCTP-7703 cells (Supporting Fig. 5, indicated by arrows), suggesting that the majority of TCTP-7703 cells had already finished cell division when Vec-7703 cells were between telophase and cytokinese.

Figure 4.

Overexpression of TCTP induces a faster M-phase exit, thereby inducing chromosome missegregation. (A) Flow cytometry was used to analyze cell-cycle distribution of Vec-7703 and TCTP-7703 cells blocked with thymidine-nocodazole (0 hours) and released for 1, 1.5, 2, or 5 hours. Percentages of cells at G0/G1, S, and G2/M are shown in (B). The value was expressed as the average of three independent experiments. (C) Flow cytometry analysis of Vec-7703 and TCTP-7703 cells treated with nocodazole for the indicated time points. The hypertetraploid population is indicated by circles. (D and E) Immunofluorescent staining of α-tubulin (green) and 4′,6-diamidino-2-phenylindole (blue) counterstaining in (D) Vec-7703 and (E) TCTP-7703 cells after being released from prometaphase arrest for 1.5 hours. Lagging chromosomes are indicated with white arrows. Red arrows indicate the formation of micro- and multinucleation in TCTP-7703 cells (1,000× magnification). (F) Percentages of mitotic cells with lagging chromosome and the formation of multi- and micronucleation are summarized in the bar chart (**P < 0.001; independent Student's t test).

We further investigated the consequences of the faster mitotic exit caused by abnormal expression of TCTP in TCTP-7703 cells. After treatment of nocodazole for 2-3 days, TCTP-7703 cells showed an increased hypertetraploid population (Fig. 4C). Furthermore, two groups of cells were stained with α-tubulin antibody at 1.5 hours after release from prometaphase. As a result, chromosomes appeared ordered and aligned on the metaphase plate, followed by completed chromosome segregation in Vec-7703 cells (Fig. 4D), whereas 56% of TCTP-7703 cells displayed abnormal mitosis (Fig. 4E). Importantly, the percentage of cells containing lagging chromosomes (indicated by white arrows) in TCTP-7703 cells (16% ± 1.6%) showed 4-fold increase, when compared to Vec-7703 cells (4% ± 0.9%) during mitosis (Fig. 4E,F). Consequently, the formation of multi- and micronucleation (indicated by red arrows) was significantly increased in TCTP-7703 cells (21% and 15%, respectively), compared to Vec-7703 cells (3% and 4%, respectively) (Fig. 4E,F).

TCTP Decreases Cdk1 Activity During Mitotic Progression.

Entry of all eukaryotic cells into mitosis is regulated by the activation of Cdk1 kinase. Dephosphorylation of Cdk1 at tyrosine 15 (Cdk1-Tyr15) in late G2 phase can induce the activation of Cdk1, thereby triggering the initiation of mitosis.16 However, decreasing Cdk1 activity in mitosis-arrested cells results in prompt mitotic exit, accompanied by defects in DNA segregation.17-19 In this study, we first studied whether TCTP could regulate Cdk1 activity during M phase. Compared to Vec-7703 cells, TCTP-7703 cells showed the higher level of Cdk1-Tyr15 at each time point after being released, suggesting that TCTP might promote M-phase exit via decreasing Cdk1 activity during mitosis progression (Fig. 5A). However, no obvious difference in Cdk1-Tyr15 level was observed at 0 hours between two groups of cells, indicating that TCTP might not affect M-phase entry. To study whether TCTP has an effect on M-phase entry, Cdc25A expression was examined in synchronized Vec-7703 and TCTP-7703 cells. No obvious difference in Cdc25A expression was detected at each time point between Vec-7703 and TCTP-7703 cells (Supporting Fig. 6).

Figure 5.

TCTP decreases Cdk1 activity through ubiquitin-proteasome degradation of Cdc25. (A) Western blotting analysis was used to compare Cdk1-Tyr15 expression between Vec-7703 and TCTP-7703 cells arrested at prometaphase (0 hours) and released for the indicated time points. Expression of protein was quantified by ImageJ software, and the ratio of Cdk1-Tyr15/β-actin was calculated and is shown in the line chart. (B and C) Expressions of Cdc25C at protein (upper panel) or messenger RNA level (lower panel) were detected in Vec-7703 and TCTP-7703 cells by western blotting analysis (B) or by reverse-transcriptase PCR (C). (D) Expression of TCTP and Cdc25C proteins in QGY-7703 cells transfected with the plasmid encoding HA-Ub, with or without TCTP expression construct, and treated with nocodazole for 6 hours in the presence or absence of MG132 (20 μM). HA-Ub and β-actin were used as a loading control. (E) Cell lysates described in (D) were immunoprecipitated with an anti-Cdc25C antibody and were then analyzed for HA-Ub and Cdc25C expression by western blotting analysis. (F) Under the treatment of CHX (50 μg/mL) for the indicated time points, Cdc25C stability during M progression was analyzed in cells released from prometaphase arrest. Expression level of Cdc25C was quantified by ImageJ software. The value is expressed as the average of two independent experiments.

TCTP Decreases Cdk1 Activity by Degrading Cdc25C via Ubiquitination.

Because dephosphorylation of Cdk1-Tyr15 (inactive form of Cdk1) is carried out by Cdc25C,16 we next investigated whether the increased level of Cdk1-Tyr15 is caused by the down-regulation of Cdc25C. As expected, Cdc25C was down-regulated at protein level, rather than mRNA level, in the synchronized TCTP-7703 cells, compared to Vec-7703 cells (Fig. 5B,C). We next verified whether TCTP is associated with Cdc25C ubiquitination during metaphase. TCTP did down-regulate Cdc25C in the absence of the proteasome inhibitor, MG132, whereas this effect could be completely abolished under MG132 treatment (Fig. 5D). Furthermore, Co-IP assay showed that overexpression of TCTP in QGY-7703 cells enhanced Cdc25C ubiquitination (Fig. 5E). To confirm the association between TCTP and Cdc25C, the stability of the endogenous Cdc25C protein was evaluated by the treatment of CHX (a protein synthesis inhibitor). During mitotic progression, TCTP-7703 cells showed an accelerated degradation of Cdc25C (Fig. 5F). These findings indicated that TCTP could degrade Cdc25C protein via an ubiquitin-proteasome pathway during mitosis, which led to the sudden drop of Cdk1 activity, followed by an accelerated mitotic exit.

Silencing TCTP Expression Inhibits Its Tumorigenicity and Abrogates Its Role in Mitotic Progression.

To further confirm whether TCTP is required for the development of HCC, short hairpin RNA (shRNA) against TCTP was used to knock down TCTP expression in HCC cell line PLC8024 (8024-shTCTP), and scrambled shRNA was used as a negative control (8024-control). Compared to 8024-control cells, TCTP knockdown in 8024-shTCTP cells (Supporting Fig. 7) caused the lower frequencies of foci formation (Fig. 6A). During the 5-week observation period, tumor formation was observed in 2 of 6 and 4 of 6 of 8024-shTCTP and 8024-control–injected nude mice, respectively (Fig. 6B). As expected, 8024-shTCTP cells showed a slower exit from M phase than 8024-control cells, particularly 2 hours after release (Fig. 6C). Meanwhile, a 3-fold decrease in hypertetraploid population was observed in 8024-shTCTP cells (5.21%), compared to 8024-control cells (16.78%) (Fig. 6D). Consistently, TCTP knockdown in 8024-shTCTP cells could increase Cdc25C level, which, in turn, increase Cdk1 activity, characterized by the lower level of Cdk1-Tyr15, compared to the control counterparts (Fig. 6E,F).

Figure 6.

Silencing TCTP expression inhibits its tumorigenicity and improper mitosis progression. (A) Representatives of foci formation induced by 8024-control and 8024-shTCTP cells. The number of foci was calculated and is summarized in the bar chart. Bars represent the mean ± SD of three independent experiments (**P < 0.01; independent Student's t test). (B) Images of tumors formed in nude mice induced by 8024-shTCTP cells and 8024-control. Average tumor volume at 5 weeks is expressed as mean ± SD for each group (n = 6) (*P < 0.05; independent Student's t test). (C) Flow cytometry was used to detect cell-cycle distribution of 8024-control and 8024-shTCTP cells treated with thymidine-nocodazole block (0 hours), then released for 2 or 6 hours. (D) The hypertetraploid population (indicated by circle) was examined in cells treated with or without nocodazole for 2 days. (E) Cdk1-Tyr15 and Cdc25C were detected by western blotting analysis in cells released from prometaphase arrest. (F) Expression levels of Cdc25C and β-actin were quantified, and the ratio of Cdc25C/β-actin was calculated for the indicated time points.

Effect of TCTP on Mitosis Contributes to Tumor Development.

To study the correlation between the tumorigenicity of TCTP and its role in mitotic progression, we isolated a single-cell population (xeno-CTL or xeno-TCTP) from xenograft tumors induced by Vec-7703 or TCTP-7703 cells and observed mitotic progression in undisturbed cells by using video time-lapse microscopy for up to 24 hours. The period for mitosis was significantly shorter (45.23 ± 4.71 minutes) in xeno-TCTP cells, compared with xeno-CTL cells (49.18 ± 2.94 minutes) (P < 0.01; Fig. 7A,B). Meanwhile, xeno-TCTP cells showed a markedly faster mitotic exit than xeno-CTL cells (Supporting Fig. 8). Moreover, xeno-TCTP cells showed higher frequencies of micro- and multinucleation, compared to xeno-CTL counterparts (Fig. 7C). In addition, cytogenetic analysis was used to compare numerical chromosomal alteration between xeno-CTL and xeno-TCTP cells. Approximately 54.2% (84 of 155) of xeno-CTL cells had 68-72 chromosomes, and the range of chromosome number was from 60 to 80. However, only 21.3% (33 of 155) of xeno-TCTP cells had 68-72 chromosomes, and xeno-TCTP cells showed a wider range (45-95) of chromosome number (Fig. 7D,E).

Figure 7.

Effect of TCTP on mitosis contributes to tumor development. (A) Time of mitosis was compared between xeno-CTL and xeno-TCTP cells (n = 20) by video time-lapse microscopy. Results summarized in (B) showed that the period for mitosis was significantly shorter in xeno-TCTP cells, compared to xeno-CTL cells (P < 0.01). (C) Representatives of micro- (indicated by arrows) and multinuclei (indicated by arrowheads) formation in Xeno-CTL and Xeno-TCTP cells. Right panels show magnified views of the boxed regions. (D) Representatives of karyotypes from Xeno-CTL and Xeno-TCTP cells. (E) Distribution of chromosome numbers was compared between Xeno-CTL and Xeno-TCTP cells. (F) Mechanistic diagram showing the effect of abnormal regulation of the CHD1L/TCTP/Cdc25C/Cdk1 pathway in HCC development. Under normal mitotic progression, Cdc25C triggers the activation of Cdk1 by the dephosphorylation of Thr14 and Tyr15 in Cdk1. The level of active Cdk1 is a key factor for maintaining the mitotic state and functions as a key switch for cell division. During HCC development, TCTP transcription is activated by a well-identified oncogene, CHD1L, which is overexpressed in over 50% of HCC cases. Overexpression of TCTP promotes the ubiquitin-proteasome degradation of Cdc25C, which leads to failure in the dephosphorylation of Cdk1 on Tyr15 and decreases Cdk1 activity. As a consequence, the sudden drop of Cdk1 activity in mitosis induces a faster mitosis exit and chromosome missegregation, which leads to the formation of lagging chromosome and micro- and multinucleation, finally causing aneuploidy and cancer development.

Discussion

Accumulation of aberrant gene expression is implicated in the progression of hepatocarcinogenesis. As a target gene of CHD1L, TCTP is highly conserved and ubiquitously expressed in various tissues, suggesting that this protein has an essential cellular function in normal cells. A recent report indicates that as a tubulin-binding protein, TCTP is temporarily associated with microtubules during G1, S, G2, and early M phases of the cell cycle and is then detached from the spindle during metaphase-anaphase transition.11 However, the underlying mechanism of TCTP overexpression in cancers and the precise mechanism by which TCTP regulates cell-cycle progression are far from clear.

In the present study, we found that CHD1L was able to bind to the 5′-upstream region (nt −733/−1027) of TCTP and could activate TCTP transcription. Clinically, expression of TCTP was found to be positively correlated with CHD1L expression in HCC samples. Furthermore, the clinical association study found that overexpression of TCTP was significantly associated with the advanced tumor stage and shorter OS time of HCC patients. More significant, TCTP was found to be an independent marker of poor prognosis. Both in vitro and in vivo functional assays demonstrated that TCTP had strong tumorigenic abilities. We found that although TCTP had no effect on G1/S transition, overexpression of TCTP in HCC cells could promote cell cycle by accelerating M-phase exit and shortening the time of cell division.

Cyclin B1/Cdk1 complex is the key regulator of mitosis in mammalian cells. In G2 phase, cyclinB1/Cdk1 accumulates in cytoplasm and on centrosomes, whereas Wee1/Myt kinases inactivate Cdk1 by phosphorylation of two residues of Cdk1: Thr14 and Tyr15.20 A key event in the activation of Cdk1 is the removal of the inhibitory phosphates. Cdc25 family members, including Cdc25A, Cdc25B, and Cdc25C, can dephosphorylate both Thr14 and Tyr15 of Cdk1. Cdc25A and Cdc25B play an important role in M-phase entry,20 whereas Cdc25C is activated in mitosis via hyperphosphorylation by cyclin B1/Cdk1. The active Cdc25C triggers the activation of cyclin B1/Cdk1 complex by the dephosphorylation of Thr14 and Tyr15 in Cdk1.21, 22 Therefore, Cdc25C plays a more prominent role in the proper execution of mitotic progression through up-regulating of Cdk1 activity. However, mitosis is an unstable cellular state and requires the continuous phosphorylation of multiple protein substrates to maintain its activation.23 The catalytic activity of Cdk1 is necessary and sufficient for maintaining the mitotic state of cells and functions as a key switch for cell division.23 The loss of Cdk1 activity is the major factor to drive the exit of cells from mitosis and ensure the correct timing of mitosis exit.19, 24-27 Interestingly, one recent study suggests that decreasing Cdk1 activity during interphase could arrest cells at G2/M phase border, whereas decreasing Cdk1 activity in mitosis causes a faster mitotic exit and premature cytokinesis.23 This finding is in agreement with our data that TCTP could down-regulate Cdk1 activity via the inhibition of the dephosphorylation of Cdk1-Tyr15, as shown by an increase in Cdk1-Tyr15 level during mitotic progression, and leads to a faster mitotic exit.

Inducible degradation of cell-cycle–regulatory proteins by the ubiquitin-proteasome pathway is one of the primary mechanisms governing passage through the cell cycle.28 Recent studies suggest that Cdc25C could be ubiquitinated during mitotic progression.29, 30 Here, we demonstrated that TCTP accelerated the ubiquitination-mediated degradation of Cdc25C during mitotic progression. In mitosis, the mitotic checkpoint is the major cell-cycle control mechanism and is also the primary defense against chromosome instability (CIN), manifested as aneuploidy, which has been strictly linked to the development of cancer. Acceleration of mitotic exit often leads to chromosomal missegregation and aneuploid progeny.31 Thus, TCTP overexpression could induce impaired chromosome segregation by increasing the formation of lagging chromosomes during mitosis and increasing the hypertetraploid population. More important, faster M-phase exit, followed by abnormal chromosome segregation, cytokinesis, and CIN, could also be observed in cell populations derived from xenograft tumors induced by TCTP-7703, suggesting the direct association between the tumorigenicity of TCTP and its effects on mitosis regulation.

In conclusion, we elucidated a novel pathway of CHD1L/TCTP/Cdc25C/Cdk1 that might play an important role in hepatocarcinogenesis by accelerating mitotic progression and inducing CIN. CIN is a hallmark of many types of human cancers, and in those tumor types where CIN is present, there exists a significant association between CIN phenotype and poor prognosis, suggesting that mitotic defects and chromosome imbalances might specifically contribute to aggressive cancer.32, 33 During HCC progression, overexpression of CHD1L is able to activate TCTP transcription, which promotes Cdc25C ubiquitination during mitotic progression and subsequently decreases Cdk1 activity, because of the failure of Cdk1 dephosphorylation on Tyr15, and accelerates mitotic exit (Fig. 7F). Further characterizing this pathway in cell-cycle progression will greatly facilitate our understanding of the HCC development and may lead to the identification of new therapeutic targets for HCC treatment.

Ancillary