Patients and liver specimens

Cells and cell culture

L02, QSG7701, PLC/PRF/5, SMMC7721, HepG2, SK-Hep-1, WRL68, Hep3B, Huh7, HCCCLM3, and HCCC97L cell lines were obtained from American Type Culture Collection (Manassas, VA) or Cell bank of Chinese Academy of Sciences (Shanghai, China). All cells were cultured in DMEM (GiBco BRL, Life Technologies) with 10% fetal bovine serum at 37°C in a humidified 5% CO₂ atmosphere.

Immunohistochemistry

Histological diagnoses of tumourous and non-tumourous formalin-fixed and paraffin-embedded tissues were confirmed on haematoxylin and eosin-stained sections. The tissue sections (3-μm thick) were dewaxed, rehydrated and then immersed in methanol containing 0.3% hydrogen peroxide for 30 min to block endogenous peroxidase activity. Subsequently, the sections were heated in a pressure cooker filled with 10 mM ethylenediaminetetracetic acid (EDTA) buffer (pH 8.0) for 2 min. After cooling, the sections were incubated in 1% blocking serum for 30 min to reduce nonspecific binding. Primary anti-ZBTB20 polyclonal antibodies[39] were diluted 1:350 and incubated with the sections at 4°C overnight. Following incubation with biotinylated secondary antisera, the streptavidin-biotin complex/horseradish peroxidase was applied. Finally, the visualization signal was developed with diaminobenzidine (DAB) and the slides were counterstained in hematoxylin.

Stained sections were evaluated in a blinded manner without prior knowledge of the clinical information using the German immunoreactive score, Immuno-Reactive-Score (IRS). Briefly, the IRS assigns sub-scores for immunoreactive distribution (0-4) and intensity (0-3), then multiplies them to yield the IRS score. The percent positivity was scored as "0" (<5%), "1" (5-25%), "2" (25-50%), "3" (50-75%,), "4" (>75%). The staining intensity was score as "0" (no staining), "1" (weakly stained), "2" (moderately stained), and "3" (strongly stained). The final ZBTB20 expression score was calculated with the value of percent positivity score plus staining intensity score, which ranged from 0 to 12 (Figure 1). We estimated IRS by averaging the values in eight fields at ×400 magnification for each specimen. Intratumoral/peritumoral ZBTB20 expression was defined as follows: low expression (score 0-6/0-3) and high expression (score > 6/4-6). Immunohistochemical analysis and scoring were performed by 2 independent investigators.

RNA extraction, cDNA synthesis and real-time RT-PCR

Total RNA was extracted from frozen tissue specimens (50-100 mg) using Trizol reagent (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's protocol. The RNA was reverse transcribed using RT-PCR kits (Applied Biosystems, Foster City, CA, USA) with an oligo d (T) primer under standard conditions. Real-time ZBTB20 PCR amplification was performed in 1× Universal Master Mix (Applied Biosystems) with gene-specific primers and probe on the ABI Prism 7300 Sequence Detection System, according to the manufacturer's instructions. The 18s rRNA levels were quantified as internal control to normalize the expression of each gene. Each reaction was repeated independently at least three times. The following primers were used: ZBTB20, sense 5'- ATGCTAGAACGGAAGAAACCCA -3', antisense 5'-TGTGAGCGTGAGAGTTTGTCA -3';18s, sense 5'-GGGAGGTAGTGACGAAAAA T-3' and antisense 5'-ACCAACAAAATAGAACCGCG-3'. Relative expression levels of genes were calculated and expressed as 2^ΔCt.

Western blotting analysis

Western blotting was performed as previously described [41]. In brief, the tissues or cells were lysed in RIPA buffer, sonicated, and protein concentrations calculated by Bicinchoninic acid (BCA) (Sigma). About 60 μg protein samples were run on a 12% SDS-PAGE gel and transferred to a polyvinylidine difluoride membrane. Membranes were blocked for 1 h in Tris-buffered saline with 0.05% Tween 20 (TBST) containing 5% skim milk. The membrane was probed with ZBTB20 antibody (dilution, 1:2000) overnight followed by washing triplely with TBST for 5 min and incubated with an IRDye 800CW-conjugated secondary antibody. After being washed three times, the membrane was detected with LI-COR imaging system (LI-COR Biosciences). Mouse anti-human β-actin antibody (dilution, 1:1000; Santa Cruz) was applied as an internal control. The signal intensity of each band on scanned autoradiograms was analysed using LI-COR imaging system. We normalised each sample to the intensity of the β-actin band, and then estimated the value as an expression index.

Follow-Up

Follow-up data were obtained for all the 102 patients with immunohistochemical analysis. The follow-up period was defined as the interval from the date of operation to the date of death or the last follow-up. Deaths from other causes were treated as censored cases. Recurrence and metastasis were diagnosed by clinical examination, AFP measurement, liver ultrasonography, and computed tomography (CT) scan. All patients were observed until December 2009, ranged from 29 to 103 months (median, 73 months). Overall survival (OS) was defined as the interval between the dates of surgery and death. Disease-free survival (DFS) was defined as the interval between the dates of surgery and recurrence; if recurrence was not diagnosed, patients were censored on the date of death or the last follow-up.

To determine factors influencing survival after operation, 13 conventional variables together with ZBTB20 expression were tested in 102 patients: age (≤50 years vs >50 years), gender, tumor size (≤50 mm vs >50 mm), serum AFP level (≤20 ng/mL vs >20 ng/mL), HBsAg (positive vs negative), Edmondson-Steiner grade (I-II vs III-IV), liver cirrhosis (yes vs no), tumor encapsulation (complete vs uncomplete vs no), vein invasion (yes vs no), tumour multiplicity (solitary vs multiple), TNM stage (I-II vs III), recurrence (yes vs no), metastasis (yes vs no), and ZBTB20 intratumoral/peritumoral protein expression level (high vs low).

Statistical analysis

A paired-samples t-test was used to compare the mRNA and protein expression of ZBTB20 between HCC tumorous and the pair-matched non-tumor tissues samples. The Pearsontest or Fisher's exact test was used to analyze the relationship between ZBTB20 expression and the clinicopathologic features. The correlation between the expression of ZBTB20 and AFP was explored using Spearman's correlation analysis. Survival curves were calculated using the Kaplan-Meier method and compared by the log-rank test. The Cox proportional-hazard regression model was used for univariate and multivariate analyses to explore the effect of the clinicopathological variables and ZBTB20 expression on survival. SPSS 16.0 software (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses and a p value < 0.05 was considered significant.

Increased Expression Levels of ZBTB20 Messenger RNA and Protein in Hepatocellular Carcinoma

Our previous work showed that ZBTB20 was developmentally up-regulated in liver and was a key repressor governing AFP gene transcription in postpartum liver, and may play an essential role in liver intrinsic functions by regulating genes such as P450 family members, glucose metabolism and the regulation of the somatotropic hormonal axis. To investigate the expression of ZBTB20 and its possible role in HCC, we first detected the protein expression in three human normal liver cell line (L02, QSG7701, WRL68) and 8 HCC cell lines (PLC/PRF/5, SMMC7721, HepG2, SK-Hep-1, Hep3B, Huh7, HCCCLM3, and HCCC97L). ZBTB20 was high expressed in most of the cells (Figure 2A, the whole image was shown in additional file 1). We further performed real-time RT-PCR for the transcriptional level of ZBTB20 from 50 frozen paired samples derived from patients who had undergone hepatectomy for histologically proven HCC. The mRNA level of ZBTB20 was significantly higher in 38 (76%) HCC tissues (tumour; T) compared with the pair-matched non-tumor tissues (NT). We calculated the copy number ratio of ZBTB20 mRNA/18s mRNA. This ratio was significantly higher in tumor samples than in NT samples (4.47 ± 2.71 versus 2.32 ± 0.64, P = 0.006) (Figure 2B).

Next, the protein level of ZBTB20 in T and NT specimens of 50 HCC cases was analyzed by western blot analysis. ZBTB20 signal was found to be positive in all T samples, and in 78% of NT samples. The intensity of each protein band was normalised against the expression of β-actin in each sample, and this ratio was treated as an expression index. Consistent with mRNA expression, the protein level of ZBTB20 in HCC tumor tissues was also significantly higher than that in NT samples (1.81-fold on average, P < 0.001) (Figure 2C).

Immunohistochemistry Analysis of ZBTB20 Expression and its Relationship with the Clinicopathological Parameters

To investigate the effect and the prognostic value of ZBTB20, immunohistochemistry was used to assess the expression of ZBTB20 in the HCC tissues sections. It was showed that ZBTB20 staining was mainly in the nucleus of tumor cells, partly in the cytoplasm. ZBTB20 expression in normal hepatocytes from para-tumoral tissue of hemangiomas of liver and cirrhotic liver were negative or weakly positive (Figure 3A,B), while mainly moderate or strong positive in HCC tumors (Figure 3C). Overall, 42 of the 102 (41.2%) cases showed high expression (IRS over 6) ZBTB20 in the tumorous tissues, while 60 (58.8%) of the cases showed low expression (IRS 0-6). Generally, ZBTB20 density was significantly higher in intratumoral tissues than in peritumoral tissues (6.37 ± 3.17 versus 2.73 ± 1.68, P < 0.001) (Table 2). Additionally, a sharp contrast was often observed between infiltrative tumorous areas of positive staining and the adjacent nontumorous area of negative staining (Figure 3D).

Our past work showed that ZBTB20 negatively regulated AFP [38]. However, in this work, no correlation between serum AFP level and tissue ZBTB20 protein expression was found. Considering the possible discrepancy of serum and liver tissue AFP level, we examined possible relationship between tissue AFP staining level and ZBTB20 expression level in HCCs, adopting Spearman's correlation analysis. In 102 samples, 61.8% cases (63/102) showed high IRS of ZBTB20 with low IRS of AFP; 29.4% cases (30/102) showed low IRS of ZBTB20 with high IRS of AFP; and 8.8% cases (9/102) showed almost the same level of IRS. As shown in Figure 4, the ZBTB20 level was negatively correlated with the staining degree of AFP in HCCs (r = -0.33, P = 0.001), while the serum AFP level had no significant correlation with the tissue AFP IRS (P= 0.538).

Correlation between ZBTB20 Expression and Prognosis

At the time of the last follow-up, 74 patients had tumor recurrence, and 59 patients had died. The 1-, 3-, and 5-year OS rates were 77.5%, 53.5%, and 45.9%, respectively; and the 1-, 3-, and 5-year DFS rates were 52.8%, 33.4%, and 26.5%, respectively. The median OS and DFS time were 18 months (95%CI, 1.06 months to 34.94 months) and 6 months (95%CI, 3.98 months to 8.01 months), respectively, for patients with high intratumoral ZBTB20 density (IRS over 6, as described in Materials and methods) and were statistically shorter than the median OS and DFS time for patients with low peritumoral ZBTB20 density [67 months (95%CI, 52.53 months to 81.47 months) and 31 months (95%CI, 13.89 months to 48.12 months), respectively (P = 0.001 and P<0.001, respectively; Figure 5A and 5B)]. Besides, patients with high peritumoral ZBTB20 had poor DFS (9 months vs 22 months, P = 0.002; Figure 5D), whereas peritumoral ZBTB20 was not associated with OS (P = 0.225; Figure 5C)