Introduction

Hepatocellular carcinoma (HCC) is one of the most common neoplasias among human populations throughout the world (Bosch et al. 1999). Risk factors such as infection with hepatitis B virus (HBV) and hepatitis C virus (HCV), dietary aflatoxin, and alcohol consumption have been associated with the development of HCC. Mutations in p53 gene, AXIN1, or CTNNB1 (encoding beta-catenin) have been reported in HCC (Bressac et al. 1991; Miyoshi et al. 1998; Satoh et al. 2000). However, the genetic events involved in hepatocarcinogenesis are still poorly understood.

Amplification of chromosomal DNA is one of the mechanisms capable of activating genes that contribute to development and progression of cancers. Earlier studies in our laboratory revealed that the 13q34 region was frequently amplified in HCCs (Yasui et al. 2002) and esophageal squamous-cell carcinomas (Shinomiya et al. 1999). Amplification of DNA at 13q34 has been observed in other types of cancer as well (Knuutila et al. 1998), indicating that the region harbors one or more protooncogenes whose amplification can lead to development or progression of a variety of tumors. In a study of primary HCCs, we identified TFDP1, which encodes transcriptional factor DP-1, as a probable target within this amplicon (Yasui et al. 2002), as TFDP1 was amplified and consequently overexpressed in some of the tumors examined. Expression levels of TFDP1 in those HCCs were closely correlated with levels of CCNE1, the gene encoding cyclin E, suggesting that overexpression of TFDP1 would lead to increased expression of this positive regulator of the G1/S transition during the cell cycle (Yasui et al. 2002).

DP-1 is a heterodimerization partner for members of the E2F family of transcription factors (E2F1 through E2F6) (Bandara et al. 1993; Girling et al.1993; Helin et al. 1993; Krek et al. 1993). Among members of the DP family, DP-1 is likely to be the major component of cellular E2F activity (Bandara et al. 1993; Wu CL et al.1995). DP-1 possesses DNA-binding and heterodimerization domains that are closely related to those of E2F proteins, but lacks a transcription-activation domain. Although DP-1 by itself has little transcriptional activity, interaction of DP-1 with E2F factors enhances DNA binding and transcriptional activity of E2F molecules (Bandara et al. 1993; Helin et al. 1993; Krek et al. 1993).

The E2F/DP-1 complex regulates expression of various cellular promoters involved in the cell cycle, in apoptosis, and in oncogenic transformation. DP-1 has shown transforming activity in cooperation with activated Ha-ras (Jooss et al. 1995) or ras plus E2F1 (Johnson et al. 1994). In transgenic mice expressing DP-1 under the control of a keratin 5 promoter, deregulated expression of DP-1 induced epidermal proliferation and enhanced carcinogenesis in epidermal tissues (Wang et al. 2001), although another type of DP-1 transgenic mice showed no obvious phenotypic changes (Holmberg et al. 1998). These lines of evidence imply oncogenic potential for DP-1. However, little is known about the actual role of DP-1 in the pathogenesis of HCC.

In the study reported here, we investigated the clinicopathological significance of up-regulated TFDP1 in HCCs. We determined expression levels of TFDP1 and E2F1, which encodes one of the best-characterized E2F factors, in primary HCCs by means of quantitative real-time reverse transcription-polymerase chain reactions (RT-PCR) and analyzed relationships between those data and various clinicopathological parameters. To assess transactivating activity of E2F1/DP-1 in HCC, we also analyzed expression levels of several known transcriptional targets of the complex. Furthermore, to determine the role of elevated TFDP1 expression in growth of HCC cells, we used antisense oligonucleotides to reduce TFDP1 expression.

Materials and methods

Cell lines and tumor samples

Hep3B, a cell line derived from a human HCC, was maintained in RPMI1640 supplemented with 10% fetal calf serum. We obtained 41 primary HCC tumors from patients undergoing surgery at the hospitals of Tokyo Medical and Dental University, and Kyoto University, Japan. All patients gave written informed consent, and all aspects of these studies were approved by the ethical committees.

Real–time quantitative RT-PCR

We quantified mRNA levels using a real-time fluorescence detection method described previously (Yasui et al. 2002). Briefly, total RNA was isolated from each HCC using Trizol (Invitrogen, Carlsbad, CA, USA), and residual genomic DNA was removed by incubating the RNA samples with RNase-free DNase I (Takara, Tokyo, Japan) prior to RT-PCR. Single-stranded complementary DNA was generated with Superscript II Reverse Transcriptase (Invitrogen) according to the manufacturer’s directions. Real-time quantitative PCR experiments were performed with an ABI Prism 7900 sequence-detection system (Applied Biosystems, Foster City, CA, USA) using SYBR Green PCR Master Mix according to the manufacturer’s protocol, using the following primers: E2F1 (F, 5’- CACAGATCCCAGCCAGTCTCTA −3’ and R, 5’- GAGAAGTCCTCCCGCACATG −3’); MYBL2 (F, 5’- TGCCAGGGAGGACAGACAAT −3’ and R, 5’- CTGTACCGATGGGCTCCTGTT −3’); CDC2 (F, 5’- AAATATAGTCAGTCTTCAGGATGTGCTTA −3’ and R, 5’- AGCCAGTTTAATTGTTCCTTTGTCAT −3’); RRM1 (F, 5’- GAAAGTGTTCAGTGATGTGATGGAA −3’ and R, 5’- TAAAGCCGAAGTAATTGTAAGAGAAATCT −3’); TYMS (F, 5’- TTTATCAAGGGATCCACAAATGCTA −3’ and R, 5’- GCCCAAGTCCCCTTCTTCTC −3’). We designed the above primers using Primer Express software (Applied Biosystems) on the basis of sequence data obtained from the NCBI database (http://www.ncbi.nlm.nih.gov/). Primers for TFDP1 (Yasui et al. 2002),TK1 (Yasui et al. 2002), DHFR (Yasui et al. 2002), CCNE1 (Muller-Tidow et al. 2001), CCNA2 (Muller-Tidow et al. 2001), PCNA (Muller-Tidow et al. 2001), and MYC (Li et al. 2000) had been described previously. GAPDH (Applied Biosystems) was used as a reference; i.e., each sample was normalized on the basis of its GAPDH content. Each assay was performed in duplicate.

Treatment of Hep3B cells with antisense TFDP1

Antisense experiments were performed as described previously (Yasui et al. 2002; Yokoi et al. 2002). Seven antisense oligonucleotides (AS11-AS17) containing phosphorothioate backbones were designed to target different regions of the primary RNA transcript of TFDP1 (NCBI database, accession no. NT_027140), according to the method proposed by Tu et al. (1998). Scrambled sequences for the antisense oligonucleotides were used as controls (SC11-SC17 respectively). Antisense oligonucleotides (400-nM each) were screened in Hep3B cells by transfection using Oligofectamine reagent (Invitrogen) according to the manufacturer’s instructions. Total RNA was isolated 24 h after transfection for real–time quantitative RT-PCR experiments to assess expression of TFDP1 mRNA. AS17, which demonstrated the greatest suppression of target molecules, was selected for further study. For subsequent assays of cell viability, Hep3B cells were seeded in 96-well plates at densities of 5.0×103 cells/well the day before transfection. Cells were transfected using 0.2 µl of Oligofectamine per well in 100 µl of Opti-MEM I Reduced Serum Medium (Invitrogen). Cells were treated with AS17 or SC17 (5’- ATGTCAGTCCCCAGAGGCCGG −3’ and 5’- GGCCGGAGACCCCTGACTGTA −3’ respectively) at final concentrations of 12.5, 25, 50, or 100 nM, or with Oligofectamine alone. Viable cells were counted by the MTT assay (cell-counting kit-8; Dojindo Laboratories, Kumamoto, Japan) 3 days after transfection, as previously described (Yokoi et al. 2002).

Statistical analysis

Statistical analysis was performed using the StatView statistical package (SAS Institute Inc., Cary, NC, USA). Chi-square tests (with Fisher’s exact probability test when necessary) were used to evaluate any association between a clinicopathological parameter and the expression level of each transcript. Any apparent associations were tested by means of a Pearson’s correlation coefficient analysis. For all statistical tests, a probability P value of less than 0.05 was considered significant.

Results

Relationships between expression levels of TFDP1 or E2F1 and clinicopathological parameters in primary HCC tumors

The mRNA levels of TFDP1 and E2F1 were determined in 41 primary HCCs using quantitative real-time RT-PCR. To clarify potential relationships between expression of those two genes and various clinicopathological parameters, we used HCCs from 37 patients whose clinical data were available. The 37 HCCs were divided into “high-expression” and “low- expression” groups at the average of the mRNA level for each gene (Fig. 1). High expression of TFDP1 significantly correlated with large (≥5 cm) tumor size (P=0.021) and with positive HBV (P=0.043) and negative HCV (P=0.016) infection (Table 1). No significant link was observed between expression level of this gene and any other parameter, such as patient’s age and gender, histological differentiation, background liver tissue, or clinical stage (TNM classification). On the other hand, expression levels of E2F1 were not correlated with any of the clinicopathological parameters examined (Table 1).

Fig. 1A, B
figure 1

Relative expression levels of TFDP1 (A) and E2F1 (B) in 37 primary hepatocellular carcinomas (HCCs). Horizontal lines indicate the average of expression levels for each gene. The average was used to determine the cutoff value for “high” and “low” expression groups

Full size image
Table 1 Relationship between expression levels of TFDP1 and E2F1 and clinicopathological parameters
Full size table

Expression of E2F/DP1 transcriptional targets

We next determined expression levels of ten genes thought to be targets of the E2F1/DP1 transcription complex (CCNE1, CCNA2, CDC2, PCNA, TYMS, RRM1, MYC, MYBL2, TK1, and DHFR) in all 41 primary HCCs and analyzed relationships between their expression levels and those of TFDP1 or E2F1 in a manner described previously (Yasui et al. 2002). Expression levels of TFDP1 correlated significantly with those of CCNE1, CDC2, PCNA, TYMS, RRM1, MYBL2, and DHFR, but not with CCNA2, MYC, or TK1 (Table 2); the closest correlation was observed between TFDP1 and CCNE1. Expression levels of E2F1 were significantly associated with those of all the presumed target genes, except MYC (Table 2).

Table 2 Correlation between expression levels of TFDP1 or E2F1 and those of their putative target genes in 41 primary hepatocellular carcinomas (HCCs)
Full size table

Inhibition of growth of HCC cells after down-regulation of TFDP1

Because elevated expression of TFDP1 was associated with large tumor size of HCCs, we down-regulated its expression in Hep3B cells, a line that had shown amplification and consequent overexpression of the gene (Yasui et al. 2002). For these experiments, we designed seven antisense oligonucleotides (AS11-AS17) targeting different regions of TFDP1. Among them, transfected AS17 exhibited the most striking reduction of TFDP1 mRNA levels (Fig. 2A). Treatment of Hep3B with AS17, but not with its scrambled control (SC17), resulted in dose-dependent inhibition of cell growth (Fig. 2B).

Fig. 2A, B
figure 2

Inhibition of growth in Hep3B cell cultures by transfection of antisense oligonucleotides targeting TFDP1. A Relative expression levels of TFDP1 mRNA determined by real-time quantitative RT-PCR. Hep3B cells were treated with 400 nM of each of seven antisense oligonucleotides (AS11-AS17), their scrambled oligonucleotides (SC11-SC17), or untreated (control). Cells were harvested 24 h after transfection. B Effect of one antisense oligonucleotide on the viability of Hep3B cells treated with the indicated concentrations of AS17 or SC17. Cell viability was determined by MTT assay 3 days after transfection

Full size image

Discussion

In the work reported here, we investigated expression levels of TFDP1 and E2F1 in primary HCCs and demonstrated that elevated expression of TFDP1, but not E2F1, was associated with large (≥5 cm) tumor size (Table 1). This finding suggests that up-regulation of TFDP1 may be a late event among the genetic alterations that occur during the course of HCC progression. On the other hand, increased expression of E2F1 may happen earlier, because it is overexpressed in most HCCs compared with their adjacent nontumorous tissues (Yasui K and Inazawa J, unpublished observation). Elevated expression of TFDP1 was also associated with positive HBV and negative HCV infection in our experiments (Table 1). Our results may be concordant with the observation that HCCs related with HBV were significantly larger in size than those with HCV (Tanabe et al. 1999).

The E2F1/DP-1 complex is thought to be involved in cell-cycle progression by regulating genes in several categories: (1) those whose products are required for DNA synthesis, e.g., thymidine kinase (TK1), thymidylate synthetase (TYMS), dihydrofolate reductase (DHFR), proliferating cell nuclear antigen (PCNA), and ribonucleotide reductase M1 polypeptide (RRM1); (2) those that encode cell-cycle regulators, e.g., cyclin A (CCNA2), cyclin E (CCNE1), and cdc 2 (CDC2); and (3) those that encode nuclear oncoproteins, e.g., c-Myc (MYC) and b-Myb (MYBL2) (DeGregori et al. 1995; Wu et al.1995). Indeed, among the transcriptional targets we examined, expression levels of TYMS, DHFR, PCNA, RRM1, CCNE1, CDC2, and MYBL2 correlated with those of both TFDP1 and E2F1 (Table 2), strongly suggesting that TFDP1 may be involved in the progression of HCC by activating those seven genes in cooperation with E2F1 or other members of E2F family.

To clarify the functional role of TFDP1 in HCC cells, we suppressed its expression by transfecting antisense oligonucleotides into an overexpressing cell line, Hep3B. AS17 reduced expression levels of TFDP1 more than SC17 (Fig. 2A). SC17 decreased TFDP1 expression, probably owing to nonspecific toxicity of oligonucleotides. Down-regulation of TFDP1 inhibited growth of these tumor-derived cells (Fig. 2B) in support of the notion that up-regulation of TFDP1 promotes growth and leads to progression of some HCCs. A larger and more definitive study will be desirable for assessing this connection, because inactivation of TFDP1 might lead to regression of HCC tumors. If so, this gene could represent an optimal molecular target for development of a novel therapy for this widespread and often intractable type of cancer.