Nine HCC cell lines (Hep3B, HepG2, HLE, HLF, HuH1, HuH2, HuH7, PLC/PRF/5 and SK-Hep1) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and were maintained as previously described (21). Primary HCC and non-cancerous tissues were collected from 151 patients who underwent liver resection for HCC at Nagoya University Hospital between January 1998 and July 2012. Clinicopathological data were collected from medical records. Specimens were classified histologically according to the 7th Edition of the Union for International Cancer Control classification (22).

Tissue samples were immediately flash-frozen in liquid nitrogen and stored at −80°C. RNA was extracted from ~5 mm² diameter tumor samples without detectable necrotic areas comprising >80% tumor cells. The corresponding non-cancerous liver tissue samples that lacked regenerative or dysplastic nodules were collected from the same patient that were >2 cm distant from the edge of the tumor. The median duration of patient follow-up was 37.9 months (range, 0.37–147 months). Postoperative follow-up included physical examinations, measurement of serum tumor markers every three months, and enhanced computed tomography scans every six months. Treatment after recurrence was generally selected from one of the options as follows: surgery, radiofrequency ablation, transcatheter arterial chemoembolization, and chemotherapy, according to tumor status and liver function. Enrollees granted written informed consent for the use of clinical samples and data as required by the Institutional Review Board of Nagoya University, Japan.

mRNA levels were determined using qRT-PCR. Total RNA (10 μg) was isolated from 9 HCC cell lines, 151 primary HCCs and adjacent non-cancerous tissues, and was used as a template for cDNA synthesis. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA (TaqMan, GAPDH control reagents; Applied Biosystems, Foster City, CA, USA) was quantified in each sample for standardization. Quantitative real-time RT-PCR was performed using the SYBR^® Green PCR Core Reagents kit (Applied Biosystems) as follows: one cycle at 95°C for 10 min, 40 cycles at 95°C for 5 sec, and 60°C for 60 sec. Real-time detection of the SYBR^® Green fluoresence was conducted using an ABI StepOnePlus™ Real-Time PCR System (Applied Biosystems). Triplicate samples of 9 HCC cell lines and 151 clinical samples were analyzed. Samples without templates served as negative controls. The expression level of each sample is shown as the value of the PDSS2 amplicon divided by that of GAPDH (23). The primer sequences are listed in Table I. PDSS2 mRNA levels were considered downregulated in tumor tissues when they were <50% compared with those of the corresponding non-cancerous tissues.

The nucleotide sequence of the PDSS2 promoter region was analyzed to determine the presence or absence of CpG islands defined as follows: at least a 200-bp region of DNA with a high HCC content (>50%) and an Observed CpG/Expected CpG ratio ≥0.6 (24,25). CpG Island Searcher software (http://cpgislands.usc.edu/) was employed to determine the locations of CpG islands (26).

PDSS2 possesses a CpG island near its promoter region, and we hypothesized that aberrant methylation regulates PDSS2 transcription in HCC. DNA samples from nine HCC cell lines treated with bisulfite were subjected to MSP. Genomic bisulfite-treated DNA from HCC cell lines was sequenced to ascertain whether the MSP amplification was reliable. The primer sequences used for MSP and bisulfite sequencing are listed in Table I. Bisulfite treatment and the sequencing procedure were performed as reported (27).

To assess the relation of promoter hypermethylation to PDSS2 transcription, HCC cells (1.5×10⁶) were treated with 5-aza-dC (Sigma-Aldrich, St. Louis, MO, USA) to inhibit DNA methylation and cultured for 6 days with medium changes on days 1, 3 and 5. RNA was extracted, and RT-PCR was performed as previously described (27).

The expression levels of hepatocyte nuclear factor 4α (HNF4α) and caudal-type homeobox transcription factor 2 (CDX2), which may associate with PDSS2 (20,28) were determined in HCC cell lines using qRT-PCR. The sequences of primers specific for HNF4α and CDX2 are listed in Table I.

IHC analysis of the localization of PDSS2 was performed using a mouse monoclonal antibody against PDSS2 (ab119768; Abcam, Cambridge, UK) diluted 1:150 in antibody diluent (Dako, Glostrup, Denmark) to probe 30 representative sections of well-preserved HCC tissue previously described (2). Staining patterns were compared between HCCs and the corresponding normal adjacent tissues. To avoid subjectivity, the specimens were randomized and coded before analysis by two independent observers who were unaware of the status of the samples. Each observer evaluated all specimens at least twice to minimize intraobserver variation (29).

Correlations between the levels of PDSS2 mRNA with those of HNF4α and CDX2 were a nalyzed using the Spearman rank correlation test. Relative levels of mRNA expression (PDSS2/GAPDH) between HCC and non-cancerous tissues were analyzed using the Mann-Whitney U test. The χ² test was used to analyze the significance of the association between the expression and methylation status of PDSS2 and clinicopathological parameters. Disease-specific and disease-free survival rates were calculated using the Kaplan-Meier method, and the difference in survival curves was analyzed using the log-rank test. We performed multivariate regression analysis using the Cox proportional hazards model to detect prognostic factors, and variables with P<0.05 were entered into the final model. All statistical analyses were performed using JMP 10 software (SAS Institute, Inc., Cary, NC, USA). P<0.05 was considered statistically significant.

A CpG island was identified in the PDSS2 promoter region (Fig. 1A), leading to the hypothesis that hypermethylation of the CpG islands regulates the expression of PDSS2 in HCC.

Significant decrease in PDSS2 mRNA levels was detected in 6 (67%) of 9 HCC cell lines compared with the mean expression level in 151 non-cancerous liver tissues (Fig. 1B). Hypermethylation of the PDSS2 promoter was detected in Hep3B, HuH2, HuH7 and SK-Hep1 cells (Fig. 1B). To determine whether hypermethylation of the PDSS2 promoter inhibited expression, PDSS2 mRNA expression levels were compared before and after treatment with the methylation inhibitor 5-aza-dC. PDSS2 mRNA levels were restored in cells with downregulated PDSS2 expression accompanying hypermethylation after 5-aza-dC treatment (Fig. 1B), indicating that promoter hypermethylation inhibited PDSS2 transcription in HCC.

The relative mRNA expression levels of PDSS2, HNF4α and CDX2 in HCC cell lines are shown in Fig. 2A. PDSS2 expression levels significantly correlated with those of HNF4α (Fig. 2B).

The mean age of the 151 patients was 64.6±9.7 years (range, 34–84 years), the male to female ratio was 5:1, and there were 37 and 84 patients with hepatitis B and C virus infections, respectively. Of the patients without HCC, 10, 87 and 54 presented with normal liver function, chronic hepatitis or cirrhosis, respectively. We classified 140 and 11 patients as Child-Pugh class A or B, respectively.

The mean level of PDSS2 mRNA compared with that of non-cancerous liver diminished gradually in the order of normal liver, chronic hepatitis and cirrhosis, indicating that chronic inflammation and fibrosis of non-cancerous liver decreased PDSS2 expression (Fig. 3A). In contrast, the type of hepatitis virus infection (hepatitis virus B, C or none) did not influence PDSS2 expression in non-cancerous liver tissues. PDSS2 mRNA expression was decreased in HCC tissues of 122 (81%) of 151 patients compared with that of the corresponding non-cancerous tissues. The mean expression level of PDSS2 mRNA was significantly lower in HCC tissues compared with that of the corresponding non-cancerous tissues (P<0.001; Fig. 3A). Moreover, poorly differentiated tumor cells expressed relatively lower levels of PDSS2 mRNA (Fig. 3B).

The expression of PDSS2 was analyzed using IHC. Representative sections with reduced PDSS2 staining in HCC tissues are shown in Fig. 3C. The overall staining intensities of 30 samples were consistent with mRNA levels detected using qRT-PCR.

Fifty-six (37%) of 151 patients were categorized with decreased PDSS2 mRNA levels in HCC tissues compared with noncancerous tissues. The disease-specific survival rate of patients with HCC with decreased PDSS2 mRNA was significantly lower compared with those without this factor (5-year survival rates, 51 and 74%, respectively, P=0.001; Fig. 4A). Decreased PDSS2 mRNA expression in patients with HCCs was significantly associated with uninvolved liver status, preoperative serum α-fetoprotein >20 ng/ml, tumor size ≥3.0 cm, tumor differentiation (moderate to poor), serosal infiltration, septum formation and advanced UICC stage (Table II). Multivariate analysis identified decreased PDSS2 mRNA expression as an independent prognostic factor (hazard ratio 2.45, 95% confidence interval 1.27–4.80, P=0.008; Table III). Patients with HCC with decreased PDSS2 mRNA levels experienced significantly earlier recurrences compared with those without (2-year recurrence-free survival rates, 38 and 59%, respectively, P=0.021; Fig. 4B). A stepwise decrease of PDSS mRNA expression in patients with HCC correlated with UICC stage (Fig. 4C).