Between 2005 and 2010, 124 patients (100 men and 24 women) with HCC were registered at the Department of Gastroenterological Surgery, of the Kumamoto University Hospital (Kumamoto, Japan). Specimens of primary HCC and adjacent normal liver tissues were obtained from the patients after written informed consent was obtained. This study was approved by the Human Ethics Review Committee of the Graduate School of Medical Sciences, Kumamoto University (Kumamoto, Japan).

Total RNA was obtained from the frozen tissue samples and cell lines using a mirVana™ miRNA Isolation kit (Ambion, Austin, TX, USA) according to the manufacturer’s instructions. Reverse transcription was performed with 1.0 μg of total RNA as previously described (20). qRT-PCR was performed on a LightCycler 480 II (Roche Diagnostics, Tokyo, Japan) using 2X PCR Master mix (Roche Diagnostics) and Universal ProbeLibrary (Roche Diagnostics). Primers were designed using the Roche website and the Universal ProbeLibrary according to the manufacturer’s instructions. The primers used were: CYLD, F: 5′-TCTATGG GGTAATCCGTTGG-3′ and R: 5′-CAGCCTGCACACTCAT CTTC-3′, and universal probe no. 83; and hypoxanthine phosphoribosyltransferase (HPRT), F: 5′-TGACCTTGATTTA TTTTGCATACC-3′ and R: 5′-CGA GCAAGACGTTCAGT CCT-3′, and universal probe no. 73. HPRT, 18S ribosomal RNA (rRNA) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were examined as the internal controls (21). HPRT was proved to be the most suitable reference gene. For amplification, an initial denaturation at 95°C for 10 min was followed by 45 cycles for 15 sec at 95°C, annealing 15 sec at 60°C, and extension 13 sec at 72°C. The experiments were performed twice to confirm reproducibility.

Paraffin-embedded tissue sections were dewaxed with xylene and rehydrated using graded concentrations of ethanol. The samples were then stained for CYLD using our previously described technique (22). Endogenous peroxidase activity was blocked using 3% hydrogen peroxide. The sections were incubated in 200X diluted primary rabbit anti-CYLD antibody (Sigma, Tokyo, Japan) overnight at 4°C. A subsequent reaction was performed with a biotin-free horseradish peroxidase enzyme-labeled polymer of the EnVision Plus detection system (Dako Co., Tokyo, Japan). A positive reaction was visualized with a 3,3′-diaminobenzidine (DAB) solution, followed by counterstaining with Mayer’s hematoxylin. Each immunohistochemical marker was independently evaluated by two blinded investigators. CYLD expression status in HCC cells was quantified as a percentage of the total number of stained cells detected in ≥5 random high-power fields (magnification, ×400) in each section. The positivity of staining cells with 10% was determined as the cut-off value.

Statistical analysis was performed using the JMP^® 8.0 software (SAS Institute., Cary, NC, USA). Values were presented as the mean ± standard deviation (SD). Differences between groups were calculated using the Wilcoxon test. P<0.05 was considered to indicate a statistically significant difference.

We performed qRT-PCR analysis in the primary HCC specimens. CYLD expression was quantified by caluculating the ratio of CYLD to HPRT1 signal. CYLD expression was detected in the tumor and non-tumor tissues. CYLD expression of tumor tissue was not markedly different compared to that of non-tumor liver tissue. For the clinicopathological evaluation, patients were allocated into two groups based on the median value of tumor-to-non-tumor (T/N) ratio of CYLD expression. Patients with a T/N ratio larger than the median T/N ratio of CYLD expression were allocated to the high expression group, while the remaining patients comprised the low expression group. Clinicopathological characteristics associated with the CYLD expression status of the 124 patients are summarized in Table I. CYLD expression was only correlated with the serum α-fetoprotein (AFP) value (P=0.0093).

The correlation between each clinicopathological characteristic and prognosis was analyzed by univariate analyses (Table II). The data indicated that poor prognosis in HCC patients correlated with tumor a diameter of >35.5 mm (P<0.0001), multiple tumors (P=0.0048), positive vascular invasion (P=0.0021), the protein induced by vitamin K absence or antagonist (PIVKA)-II >108 (P=0.0278), and low CYLD expression (P=0.0406) (Fig. 1A). In the multivariate analysis, CYLD expression was not an independent factor for predicting poor prognosis (data not shown). Although CYLD expression was not significantly correlated with disease-free survival (P=0.1021) (Fig. 1B), the low CYLD expression group had more patients with early recurrence within 2 years (30/37 patients) compared to the high CYLD expression group (17/31 patients; P=0.016).

Among 70 HCC cases, 53 (75.7%) were positive for CYLD expression. CYLD expression was heterogeneously distributed in the tumor tissue and downregulated in tumor cells. In Fig. 2A, a representative case of HCC shows that a number of tumor cells (T1) with a high CYLD expression are well-differentiated and that they demonstrate a trabecular pattern. Conversely, other tumor cells (T2) with low CYLD expression lost their cell polarity and demonstrated dense chromatin in the nucleus. Another case of HCC comprising tumor cells with dense chromatin and a small nucleus that lost CYLD expression, despite being surrounded by CYLD-expressing tumor cells with more cytoplasm and only faint chromatin in the nucleus (Fig. 2B). However, CYLD protein expression was not associated with tumor-related factors, such as tumor size, tumor diameter, vascular invasion, tumor differentiation and prognosis (data not shown). To confirm the correlation of CYLD-mRNA expression with protein expression, CYLD-mRNA expression normalized by HPRT-mRNA expression in tumor tissue was compared between the high and low-CYLD protein expression groups. This finding showed that the high-CYLD protein expression group demonstrated a markedly higher CYLD-mRNA expression compared to the low-CYLD protein expression group (P=0.036) (Fig. 2C).