Potential conflict of interest: Nothing to report.
The pituitary tumor transforming gene 1 (PTTG1) protein is cell-cycle regulated and is identified as a human securin that inhibits sister chromatid separation and is involved in transformation and tumorigenesis. PTTG1 has very low or undetectable expression in most normal human tissues, but it is abundantly expressed in malignant cell lines and pituitary tumors. In this study, we investigated human PTTG1 expression in 62 hepatocellular carcinoma (HCC) specimens using quantitative real-time reverse transcription polymerase chain reaction analysis. We found that, compared with corresponding noncancerous liver tissues, PTTG1 was remarkably overexpressed in HCCs (PTTG1/β-actin; 0.443 ± 0.073 vs. 0.068 ± 0.007; P < .0001). Furthermore, we found a significant correlation between PTTG1 expression and serum alpha-fetoprotein level (P < .001). Univariate and multivariate analyses revealed that the PTTG1 messenger RNA (mRNA) expression was an independent prognostic factor for disease-free (odds ratio 2.70; P = .037) and overall (odds ratio 5.35; P = .007) survival. Moreover, we discovered a significant relationship between PTTG1 expression and intratumoral microvessel density. Our data supported an important role for PTTG1-mediated upregulation of fibroblast growth factor (FGF)–2, one of angiogenesis and modulation of tumor progression, in hepatocarcinogenesis. In conclusion, PTTG1 might be critically involved in the development of HCCs through the promotion of angiogenesis. PTTG1 is overexpressed in HCC and our results suggest that PTTG1 mRNA expression has prognostic significance for the survival of postoperative patients with HCC. (HEPATOLOGY 2006.)
Hepatocellular carcinoma (HCC) is a common cancer worldwide, especially in Southeast Asia and Africa, and it is the third leading cause of death from malignant disease in Japan.1 Surgical treatments including hepatic resection and liver transplantation improve the chances of survival for a patient with HCC. However, most patients experience recurrence or metastasis after surgery, and their prognosis remains poor.2 Therefore, to further improve patient survival, identifying genetic markers of recurrence or metastasis of HCC is important.
Recent developments in microarray technology have ushered in a new era in medical science, allowing precise identification of the genetic factors associated with various tumors.3–5 Ramaswamy et al.6 reported that 17 genes were associated with metastasis in several primary solid tumors and that the pituitary tumor transforming gene 1 (PTTG1) protein was one of them. PTTG1 was originally isolated from cloned rat GH4 pituitary tumor cells,7 and its human homolog hpttg was cloned from a thymus cDNA library.8 It is identified as a human securin that inhibits sister chromatid separation and is involved in malignant transformation and tumorigenesis.9 At the end of metaphase, securin is degraded by an anaphase promoting complex, releasing tonic inhibition of separin, which in turn mediates the degradation of cohesions, the proteins that hold sister chromatid together. Overexpression of the undegradable PTTG1 disrupts the sister chromatid separation, generates chromosomal instability, and thereby increases cell susceptibility to acquisition of further mutations during subsequent divisions.10, 11 Moreover, p53 specifically interacts with securin both in vitro and in vivo and securin inhibits the ability of p53 to induce cell death. This activity demonstrated its oncogenic potential.12, 13
PTTG1 overexpression causes stimulated expression and secretion of fibroblast growth factor (FGF)-214 and vascular endothelial growth factor (VEGF)15 and is associated with an angiogenic phenotype in human tumors.16 Through the stimulation of FGF-2 production and secretion, PTTG1 may enhance tumor growth and progression.
In this study, we investigated human PTTG1 expression in 62 HCC specimens using real-time quantitative reverse transcription polymerase chain reaction (RT-PCR). We elucidated that PTTG1 messenger RNA (mRNA) was overexpressed in these tumors when compared with the corresponding noncancerous liver tissues, and we found that there was a significant positive correlation between PTTG1 expression and FGF-2 expression. Moreover, we found that PTTG1 mRNA expression was significantly correlated with intratumoral microvessel density (MVD). Our results suggest that PTTG1 mRNA expression has prognostic significance for the survival of patients with HCC and may be critically involved in HCC development through the promotion of angiogenesis.
Cancerous tissues and surrounding non-cancerous hepatic parenchyma were obtained from 62 primary HCC patients who underwent resection at Nagoya University Hospital, Japan, from August 1989 to December 1997. Samples were from 53 men and 9 women of aged 21 to 77 years (mean, 61 years). The average follow-up period in the prognosis study was 4.4 ± 0.3 (mean ± SD) years. Tumor sizes ranged from 1.5 to 14.0 cm, with a mean size of 4.2 ± 2.8 cm. Specimens were histologically classified by the Japanese staging system of the Liver Cancer Study Group of Japan.17 In the corresponding noncancerous parenchyma, cirrhosis was found in 33 patients (53.2%), and tissue without cirrhosis was found in 29 patients (46.8%). The human HCC cell lines HepG2, Hep3B, HLE, HLF, HuH1, HuH2, HuH7, and PLC/PRF/5 were obtained from American Type Culture Collection. The ethics committee of the hospital approved this study. Informed consent was obtained from every patient before the study was started for the use of his or her resected tissues.
RNA Extraction and cDNA Synthesis.
Total RNA from the cell lines was isolated using RNeasy Mini Kits (Qiagen, Hilden, Germany), according to the manufacturer's protocol. Specimens were taken from viable and non-fibrotic areas of tumors and noncancerous tissues. They were quickly frozen in liquid nitrogen and then stored at −80°C until use. Extraction of RNA from tissues was conducted using a CsCl density-inclination ultra-centrifugation technique. A 27-μL aliquot of solution, including 10 μg total RNA and 10 μg random primer (Roche, Mannheim, Germany), was incubated at 70°C for 10 minutes. After cooling on ice, 23 μL reverse transcription (RT)-enzyme mixture was added into the solution. The final 50 μL RT-enzyme mixture contained 10 μL 5× RT buffer, 5 μL 0.1 mol/L dithiothreitol, 3 μL 10 mmol/L dNTP mixture, 200 U RNA guard RNase inhibitor, and 1,000 U Superscript II reverse transcriptase (Life Technologies, Tokyo, Japan). RT reactions were carried out at 42°C for 90 minutes and then at 90°C for 2 minutes.
Real-Time Quantitative RT-PCR Analysis.
The primer sequences are listed in Table 1 for real-time quantitative RT-PCR analysis. All PCR reactions were performed with the SYBR Green PCR Core Reagents kit (Perkin-Elmer Applied Biosystems, Foster City, CA) under the following conditions: 1 cycle at 50°C for 2 minutes, 1 cycle at 95°C for 10 minutes, then 45 cycles at 95°C for 15 seconds and at 60°C for 1 minute. Real-time detection of the SYBR Green emission intensity was conducted with an ABI prism 7000 Sequence Detector (Perkin-Elmer Applied Biosystems). An equivalent amount of cDNA sample, which was derived from 40 ng total RNA, was used for each PCR reaction. External standards were prepared by serial dilution (1:2 to 1:10,240) of cDNA from the HepG2 HCC cell line. The mRNA in each sample was then automatically quantitated with reference to the standard curve constructed each time using ABI 7000 software. Quantitative RT-PCR was performed at least 3 times, including a no-template sample as a negative control. To standardize the amount of RNA, we quantified the expression of β-actin in each sample and then divide the amounts of expression of PTTG1, FGF-2, and VEGF by that of β-actin.
Table 1. Oligonucleotide Sequences of PCR Primers
Western Blot Analysis.
Cells and tissues were retrieved into RIPA buffer (50 mmol/L Tris, pH 8.0, 150 mmol/L NaCl, 1% NP-40, 0.1% sodium dodecyl sulphate (SDS), and 0.5% deoxycholate) containing 10 μg/mL aprotinin, 10 μg/ml leupeptin, 2 mmol/L ethylenediamine tetraacetic acid, and 1 mmol/L phenylmethylsulphonyl fluoride. The protein concentration was determined using a BCA Protein Assay Kit (Pierce, Rockford, IL). Protein (20 μg) was resuspended in SDS sample buffer (60 mmol/L Tris-HCl, a pH of 6.8, 0.05% bromophenol blue, 5% glycerol, 4% SDS) containing 5% β-mercaptoethanol and then heated at 70°C for 10 minutes. The protein was then subjected to a 4% to 12% gradient Bis-Tris Gel (Invitrogen, Carlsbad, CA) and electrotransferred to a Hybond-enhanced chemiluminescence nitrocellulose membrane (Amersham Pharmacia Biotech, Buckinghamshire, UK). The membrane was incubated overnight at 4°C with an anti-human securin monoclonal antibody (diluted 1:100, Novocastra Laboratories, Newcastle, UK) and subsequently incubated with a horseradish peroxidase–conjugated secondary antibody (1:500) for 1 hour at room temperature. The reaction was detected using an ECL system (Amersham Pharmacia Biotech). Proteins were then re-blotted using anti–β-actin (diluted 1:700, Sigma, St Louis, MO) as an internal control.
Immunohistochemical Staining and Microvessel Counting.
An immunohistochemical analysis was performed on paraffin-embedded sections using the Envision kit (DAKO, Glostrup, Denmark) following the manufacturer's instructions. The sections were autoclaved for 10 minutes at 121°C for antigen retrieval. Anti-CD34 monoclonal antibodies (DAKO) were applied to the sections at a dilution of 1:100 as the primary antibodies. The section slides were counterstained with hematoxylin; they were then dehydrated and mounted. For microvessel counting, the four most highly vascularized areas were counted in 200× fields (0.74 mm2 per field), and the average counts were then recorded.
Anti-PTTG1 polyclonal antibodies (Zymed Laboratory Inc., South San Francisco, CA) were applied to the sections at a dilution of 1:50. The presence of staining was evaluated by a single pathologist (Y.Y.) according to the overall level of the immunostaining.
The relative mRNA expression levels (PTTG1/β-actin) were calculated from the quantified data. An unpaired t test was used to analyze the differences in the PTTG1 expression levels between HCCs and the corresponding noncancerous hepatic tissues. To analyze the correlation between PTTG1 expression and clinicopathological parameters, differences in the numerical data between the two groups were evaluated using the Mann-Whitney U test. Differences in the numerical data among more than three groups were evaluated using the Kruskal-Wallis test. Overall and disease-free survival rates were then calculated using the Kaplan-Meier method, and the differences in survival curves were analyzed using the log-rank test. Survival was censored if the patient was still alive or had died of other causes. Independent prognostic factors were analyzed by the Cox proportional hazards regression model in a stepwise manner. Correlations among different mRNA expression levels were analyzed using the Pearson rank sum test. All the statistical analyses were performed using Stat View (Version 5.0) software (Abacus Concepts, Berkeley, CA). Data are expressed as mean ± SE. P < .05 denoted the presence of a statistically significant difference.
PTTG1 mRNA Expression in HCC.
We examined 62 HCC samples and corresponding noncancerous hepatic tissues for PTTG1 mRNA expression using real-time quantitative RT-PCR. Most HCCs had a noticeable upregulation of PTTG1 expressions, as shown in Fig. 1A. In contrast, the noncancerous liver tissues had low PTTG1 expression levels. The average PTTG1/β-actin level in HCCs was significantly higher than that in noncancerous livers (Fig. 1B; 0.443 ± 0.073 vs. 0.068 ± 0.007; P < .0001).
Correlation Between PTTG1 mRNA Expression and Clinicopathological Parameters in HCC.
The correlation between PTTG1 expression and clinicopathological parameters in patients with HCC was statistically analyzed. The results are listed in Table 2. There was a significant correlation between the high expression of PTTG1 and high serum alpha-fetoprotein (AFP) levels (P < .001). PTTG1 was overexpressed more often in HCC with capsular infiltration than in HCC without capsular infiltration; it was also overexpressed more often in HCC with vascular invasion than in HCC without vascular invasion. However, the differences were not statistically significant. No significant correlation was found between PTTG1 expression and other variables, such as age, sex, tumor size, tumor multiplicity, septal formation, and capsular formation.
Table 2. The Correlations Between PTTG1 Expression and Clinicopathological Parameters in 62 Patients With HCC
Univariate and Multivariate Analysis of Prognostic Factors for Patients With HCC.
To quantify the upregulation of PTTG1 expression in HCC, we calculated the ratio of the expression level in the tumor to that in the corresponding noncancerous hepatic tissues (T/N ratio; PTTG1/β-actin in T divided by PTTG1/β-actin in N). For statistical analysis of the PTTG1 expression, the specimens were divided into two groups: 31 (50.0%) high expressers and 31 (50.0%) low expressers, using a cutoff level of 5.020, which was the median value of the T/N ratio. We analyzed the disease-free survival rates and overall survival rates to assess the prognostic significance of PTTG1 expression. The 5-year disease-free and overall survival rates of the 62 patients with HCC were 50.1% and 70.8%, respectively. Using Kaplan-Meier curve assessment, we found that patients having higher PTTG1 expression levels had lower 5-year survival rates than the patients having lower PTTG1 expression levels (disease-free survival rate, 35.1% vs. 64.7%; P = .019; overall survival rate, 58.6% vs. 82.3%; P < .01; as shown in Fig. 2).
To evaluate the potential of using PTTG1 expression in determining the postoperative prognosis of HCC patients, univariate analysis using a Cox proportional hazards regression model was conducted. The results showed that both PTTG1 mRNA level and vascular invasion were significant predictors of tumor recurrence (P = .024 and P = .029, respectively). However, only the PTTG1 mRNA level was a significant predictor of overall survival (P = .017) (data not shown). Multivariate analysis was conducted with five other prime variables (liver cirrhosis, tumor size, tumor multiplicity, serum AFP index, and pathological TNM stage) that had been regarded as prognostic factors for patients with HCC in previous studies.1, 18 A high PTTG1 mRNA level was still the strongest variable for independently predicting disease-free and overall survival (P = .036 and P = .007, respectively; Table 3). The odds ratios of higher PTTG1 mRNA levels as compared with the lower PTTG1 levels were 2.75 and 5.15 (95% confidence interval, 0.142-0.934 and 0.058-0.645, respectively).
Table 3. Multivariate Analysis of Disease-Free and Overall Patient Survivals With Hepatocellular Carcinoma
95% Confidence Interval
95% Confidence Interval
Abbreviation: AFP, alpha-fetoprotein.
PTTG1 expression (High:Low)
AFP (<70 ng/mL: ≥70 ng/mL)
Tumor multiplicity (solitary: multiple)
Tumor size (<3.5 cm: ≥3.5cm)
Vascular invasion (absent: present)
Liver cirrhosis (absent: present)
Pathological stage (I and II: III and IV)
PTTG1 Expression at the Protein Level.
To investigate the PTTG1 expression at the protein level, Western blot analysis was conducted in a HepG2 cell line and five representative HCC tissues. Different PTTG1 expression levels were revealed (Fig. 3A). PTTG1 was expressed at a high level in HepG2 and cancerous tissues, but only at a very low level in noncancerous tissues. The results from the Western blot analyses tended to parallel the results of relative mRNA expression examined using quantitative RT-PCR (Fig. 3B).
An immunohistochemical study using anti-PTTG1 polyclonal antibodies was performed to determine whether PTTG1 is also expressed in HCC specimens. PTTG1 protein was detectable in cancer cells despite non-expression in corresponding noncancerous tissues (Fig. 4A-B). With regard to subcellular localization, PTTG1 staining was observed in the cytoplasm of tumor cells.
FGF-2, VEGF, and PTTG1 Expression in HCC.
We also examined the cancerous tissues for mRNA expression of FGF-2 and VEGF using real-time quantitative RT-PCR. FGF-2 is known to regulate the endothelial expression of VEGF, and in this study, we found a significant correlation between the mRNA expression of FGF-2 and that of VEGF (Fig. 5A). Furthermore, we investigated the correlation between the mRNA expressions of these growth factors with that of PTTG1. We found a highly significant positive correlation between the mRNA expression of PTTG1 and that of FGF-2 in HCC (Fig. 5B; P < .0001, r = 0.78). A significant positive association between mRNA expression of PTTG1 and that of VEGF was also observed (Fig. 5C; P = .0001, r = 0.47).
Relationship Between PTTG1 Expression and the Intratumoral MVD.
Specific staining of microvessels by anti-CD34 was observed in all tumor specimens (Fig. 6A). The median tumor MVD was 155/field (range, 110-246). Significant correlations between intratumoral MVD and the mRNA expression of both angiogenic growth factors, FGF-2 and VEGF, were observed (Fig. 6B-C). Furthermore, we found a highly significant positive correlation (Fig. 6D; P < .0001, r = 0.80) between mRNA expression of PTTG1 and the MVD in HCC.
A large number of patients who undergo hepatectomy for HCC are reported to develop new tumors in the residual liver.2 To achieve adequate and effective management, assessing the aggressiveness or malignancy of this neoplasm is important. However, although a variety of biological markers have been investigated, many remain insufficient.
In this study, we analyzed PTTG1 mRNA expression by quantitative RT-PCR and obtained the evidence that PTTG1 mRNA was significantly overexpressed in human HCC compared with the corresponding normal liver tissue. We also investigated the correlation between PTTG1 expression and clinicopathological parameters in HCC. Our data indicated that PTTG1 mRNA expression was not only noticeably upregulated in tumors but also associated with serum AFP level (Table 2). Serum AFP level was previously reported to be a significant prognostic factor.1 No significant correlation was seen between PTTG1 expression and other aggressive phenotypes, including tumor multiplicity, histological type of tumor, tumor size, or pathological TNM (tumor, node, metastasis) stage in our cohort. The vascular invasion and capsular infiltration, which were reported as the prognosticators,18–20 are not also significantly correlated, but tend to be more frequent in patients with higher PTTG1 expressions than in patients with lower PTTG1 expressions. Further examination is necessary to explain how the different PTTG1 mRNA expressions are associated with tumor invasiveness.
FGF-2 and VEGF show a synergistic effect and intratumoral MVD and the expression of these angiogenic factors were significantly correlated in HCCs.21, 22 In this study, we confirmed correlative expression of these factors and revealed the significant correlation between MVD and the expression of both FGF-2 and VEGF. Moreover, previous studies reported that seemingly direct links among PTTG1, FGF-2, VEGF, and angiogenesis were observed in pituitary tumors,14–16, 23 and an association between tumor vascularity and increased PTTG expression was recognized.16 Motivated by these reports, we compared the expression of these angiogenic factors with that of PTTG1 in our cohort of HCCs. We found a highly significant positive correlation between the mRNA expression of PTTG1 and that of FGF-2, and a highly significant positive association between PTTG1 expression and intratumoral MVD was also observed. These results suggested that PTTG1 might be involved in tumoral angiogenesis through stimulation of these angiogenic growth factors and enhanced tumor growth and progression in HCCs. Because PTTG1 disrupts sister chromatid separation, previous reports concluded that the overexpression of this gene generates chromosomal instability.10, 11 A more recent report, however, demonstrated that high expression of PTTG1 did not contribute to the development of DNA aneuploidy in myeloid leukemias.24
Univariate and multivariate analyses showed that only PTTG1 expression was significantly associated with both disease-free and overall survival rates. This result indicates that PTTG1 overexpression might help to identify the HCC patients with poor prognoses and that its expression could be a potentially novel prognostic marker of HCC. A recent study found that there was no correlation between PTTG1 and proliferating cell nuclear antigen expression in thyroid cancer, suggesting that the observed increase in the PTTG1 expression in cancer is not a marker of increased cell turnover but reflects an involvement of PTTG1 in some other fundamental pathway.25 Previous studies on PTTG1 overexpression in colorectal, esophagus, breast, thyroid, and gastric cancers indicated that its expression was correlated with vascular formation and lymph node metastases.25–30 As mentioned previously, our results indicated that a high PTTG1 expression might reflect the malignant potential of HCC, probably through tumor angiogenesis. A clinical significance of MVD has been demonstrated in various neoplasms. The prognostic value of angiogenesis in HCC is controversial, with some recent reports demonstrating that MVD was predictive of early recurrence in patients with HCC.31–33 The detailed mechanism by which PTTG1 overexpression contributes to angiogenesis in HCC and the poor prognosis of patients with HCC must be elucidated by further study.
In conclusion, PTTG1 is overexpressed in HCC and that PTTG1 mRNA expression is associated with tumor angiogenesis and is significantly associated with disease-free and overall survival rates. Detection of high PTTG1 expression in surgically excised HCC tissues might help to identify patients with aggressive disease who will need adjuvant therapy. Our results raise the possibility that unraveling treatment strategies aimed at functional abrogation of PTTG1 could provide new therapeutic approaches in the management of HCC.
The authors thank Y. Nishikawa, A. Tagashira, and H. Mizuno for their excellent technical assistance.