We collected tumors from 85 consecutive patients who had undergone surgery for HCC at the Jinan Military General Hospital from January 2006 to September 2010. The Ethics Committee of the Jinan Military General Hospital approved the protocol of this study. Among the 85 patients, 55 had serum α-fetoprotein (AFP) ≥ 30 μg/L, and 73 were sera positive for hepatitis B surface antigen (HBsAg). On gross examination, 3 cases had tumor sizes that were ≤ 2 cm, and 82 had tumor sizes > 2 cm (median tumor size, 6.1 cm; range, 1.0-16 cm). Histopathological diagnoses were made according to the pathological classification system of the World Health Organization (2000), and the tumor was staged following the tumor-node-metastasis classification of the International Union Against Cancer[33]. Nine cases were well differentiated; 60 cases were moderately differentiated; and 16 cases were poorly differentiated. In total, 56 HCC cases had liver cirrhosis; 25 cases had chronic hepatitis; and 4 cases had basically normal liver tissue. We also collected 3 cases of lung metastasis. Furthermore, tissues of comparative normal liver obtained during surgery for liver cholelithiasis (n = 3) and HBV-infected liver biopsies (n = 5) were studied.

Nineteen of the 85 cases included chronic (n = 8) and cirrhotic (n = 11) HCC. From these, fresh tissues were obtained immediately after resection, including HCC tumor and adjacent nontumorous liver tissues. In addition, normal liver tissues (n = 3) with no HBV infection were obtained during surgery for liver cholelithiasis. In these 22 cases, one portion of the fresh tissue was snap frozen in liquid nitrogen immediately and stored at -80 °C; the remainder portion was fixed in 10% buffered formalin and embedded in paraffin.

The available patient clinicopathological information included gender, age, serum AFP, serum HBsAg, tumor size, tumor stage, histological grade and cancer-specific survival time.

Total RNA was extracted from 10-mm frozen HCC tissue sections. To isolate the RNA from defined areas containing ≥ 80% tumor cells, all tumors were manually microdissected under direct visual control through a dissecting microscope. Total RNA in the frozen tissues was extracted using Trizol (Invitrogen) following the manufacturer’s recommendations. Total RNA was digested with DNase I (Invitrogen) and was used for the first-strand cDNA reaction. The reaction mixture consisted of 5 μg of DNase I-treated RNA, 1 × reverse transcriptase buffer, 2.5 mmol dNTP mix, 3.5 μmol oligo primer, and 2.5 U/mL MultiScribe™ reverse transcriptase (PE Applied Biosystems). Each sample was handled using the same protocol, with the exception that reverse transcriptase was added to exclude the presence of interference from genomic DNA.

Real-time reverse transcription polymerase chain reaction (qRT-PCR) was carried out using SYBER green dye in a Rotor Gene 3000 Detection System (Corbett Research, Sydney, Australia). Each SYBER green reaction (25 μL) contained one microliter diluted cDNA and 10.5 μL SYBR Green PCR Master Mix, as well as 5 pmol forward and reverse primer (Ror2: forward 5’-AGGTGCCTATGCAAGTTCA-3’, reverse 5’-TGTGCGAGGTTTAAGGTCTA-3’). Samples were activated by incubation at 95 °C for 5 min and denatured at 95 °C for 20 s. This was followed by annealing at 60 °C for 20 s and extension at 72 °C for 20 s, for 40 cycles. In all of the cDNA samples, gene expressions of Ror2 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (forward 5’-GAAGGTGAAGGTCGGAGTC-3’; reverse 5’-GAAGATGGTGATGGGATTTC-3’), an internal quantitative control, were determined by SYBR green fluorescence using the Rotor-Gene 3000; the ratios of Ror2 to the housekeeping gene GAPDH represented the normalized relative levels of Ror2 expression. A non- template negative control was also included in each experiment. Analyses of all tumor samples were carried out at least twice, and the mean value was calculated.

Immunohistochemical staining was performed on thin sections (4 μm) of paraffin-embedded archival tissue. The samples were dewaxed with xylene/ethanol before antigen retrieval (i.e., pressure cooked for one minute at full pressure, 15 psi, in 0.001 mol/L EDTA buffer, pH 8.0). The primary antibodies used were: Wnt5a (LS-C47384, Lifespan, 1:200), Ror2 (PAB3386, Abnova, 1:200), β-catenin (C19220, BD Transduction Laboratories, 1:400) and Ki-67 (MIB-1, Dako, 1:100). Immuno- histochemical staining of antibodies was done using the Dako Envision Plus System (K5007, Dako). The antibody binding was visualized with 3, 3’-diaminobenzidine tetrahydrochloride (DAB) before brief counterstaining with Mayer’s hematoxylin. For monoclonal antibodies of mouse origin, negative controls were obtained using isotypic mouse immunoglobulin in the same dilution as the primary antibody of concern. All control experiments gave negative results.

Two authors (Cao YC and Jiang H) who had no knowledge of the patients’ clinical status reviewed all of the immunostained sections. Cases with discrepant results were re-evaluated jointly until agreement was reached. For expression of Wnt5a, Ror2 and β-catenin protein, in cases with multiple areas of low intensity that occurred during evaluation of immunostaining, five areas were selected at random and scored.

The degree of immunohistochemical staining was recorded using a semi- quantitative and subjective grading system that considered both the intensity of staining and the proportion of tumor cells that had an unequivocal positive reaction. Grades for stain intensity were: 0: No staining; 1: Weak staining; 2: Positive staining; and 3: Strong staining. For rating stained areas: 0: No staining; 1: Positive staining in < 10% of tumor cells; 2: Positive staining in 10% to 50% of tumor cells; 3: Positive staining in > 50% of tumor cells. The staining index was calculated as the staining intensity multiplied by the positive area.

Ki-67-positive cells were counted by viewing ≥ 200 HCC cells from ≥ 10 randomly selected fields. The percentage of antigen-positive nuclei among the total number of nuclei counted was calculated to obtain the nuclear labeling index (LI).

In the subsequent statistical analysis, the cutoff points for the staining index categories were mainly based on median values, as well as each marker’s frequency distribution curve and the size of the subgroups. Therefore, cytoplasmic Wnt5a and Ror2 and membranous β-catenin staining indices were categorized by their median value as high (> 4) or low (0-4), and the cytoplasmic β-catenin staining index was categorized as high (> 3) or low (0-3). However, nuclear β-catenin expression was categorized based on the absence (staining index = 0) or presence (staining index ≥ 1) of staining. The Ki-67 labeling indices were divided into two groups (LI < 10% and LI ≥ 10%).

To determine the prognostic factor, the outcome of the 82 patients was determined by reviewing their medical charts. The follow-up period ranged from one to 54 mo (average: 31.3 mo; median: 27.0 mo). The end point in the analysis was HCC-related death. The overall and disease-free survival rates were estimated using the Kaplan-Meier method and compared with the log-rank test. The prognostic analysis was carried out with univariate and multivariate Cox regressions models.

The differences in Ror2 mRNA expression between HCC and nontumorous liver tissue was statistically analyzed using Student’s t-test and one-way analysis of variance (ANOVA) for multiple comparisons. The correlations between the clinicopathological parameters and Wnt5a, Ror2 and β-catenin protein expression were analyzed using the χ² or Fisher’s exact tests.

Pearson’s correlation was used to determine the correlation between mRNA and protein expression, as well as between the expressions of different proteins. All statistical calculations were carried out using SPSS software (for Windows, version 13.0). A significant difference was defined at P < 0.05.

The Ror2 gene (mRNA) expression levels relative to that of GAPHD in normal, HCC, chronic hepatitis, cirrhotic liver and adjacent nontumorous liver tissues are shown in Figure 1. Ror2 mRNA levels were elevated in chronic hepatitis (5.420 ± 5.492, n = 11) and cirrhotic liver tissues (6.128 ± 5.252, n = 8) compared to that of normal (3.381 ± 1.182, n = 3) and HCC (3.189 ± 3.856, n = 19). Based on Student’s t-test, statistically significant differences were found between the Ror2 mRNA levels in HCC vs adjacent nontumorous (chronic hepatitis or cirrhotic) liver tissues (P = 0.029), but not between HCC and normal liver tissues (P = 0.934) or normal and adjacent nontumorous liver tissues (P = 0.094). The Ror2 mRNA level in moderately and poorly differentiated tumor tissues (n = 16) was greater by 7.2-fold (P = 0.014) than the level in well-differentiated tumor tissues (n = 3). No significant differences were found between Ror2 gene expression levels and other clinicopathological findings such as age, serum AFP concentration, tumor size, and HCC tumor stage.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Real-time reverse transcription-polymerase chain reaction analysis of Ror2 gene (mRNA) expression in hepatocellular carcinoma, chronic hepatitis, cirrhotic and normal liver tissue. Ror2: Receptor tyrosine kinase-like orphan receptor 2; HCC: Hepatocellular carcinoma; GAPDH: Glyceraldehyde-3-phosphate dehydrogenase; Bars: mean; Columns: SD; N: Normal; CH: Chronic hepatitis; CC: Cirrhosis; M, P, W: Moderately, poorly, and well differentiated tumor tissues, respectively.

Immunohistochemistry was performed to evaluate Ror2 protein expression in tumor and non-tumorous liver cells. In non-tumorous liver cells and HCC tumor cells, Ror2 protein expression was displayed in the cytoplasm, but in stromal cells Ror2 protein was not observed. In comparative normal liver cells Ror2 was negative or weakly expressed (Figure 2A and B), whereas all chronic hepatitis, cirrhotic, and dysplastic liver cells exhibited positive immunostaining for Ror2 (Figure 2C and D), In 62/85 (72.9%) of the HCCs, Ror2 immunostaining was reduced or absent (Figure 2E and F).

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Immunohistochemical staining for Ror2 in hepatocellular carcinoma. Patient-matched normal liver cells showed negative (A) or weak expression (B) of receptor tyrosine kinase-like orphan receptor 2 (Ror2). Liver cirrhosis cells (C) and dysplastic liver cells (D) exhibited strong positive immunostaining for Ror2. Tumor cells (E, F) showing negative Ror2 staining in hepatocytes, while strong cytoplasmic staining is seen in adjacent nontumorous cells. Original magnification, 200 × in A; 400 × in B and D; 100 × in C.

A significant correlation was found between the normalized Ror2 gene expression ratio and the protein expression level in normal, tumor and non-tumorous liver tissues (r = 0.254, P = 0.021). Furthermore, statistical comparisons between Ror2 mRNA expression and patients’ clinicopathological features revealed a significant negative association between Ror2 mRNA and tumor stage (P < 0.001), and between Ror2 mRNA and serum AFP (P < 0.001). However, there were no significant differences between Ror2 protein expression and the other clinicopathological findings in HCC (Table 1).

Wnt5a protein expression was observed in the cytoplasm of non-tumorous liver and tumor cells, but nowhere in stromal cells. There was little or no Wnt5a seen in normal liver cells. However, all chronic hepatitis, cirrhosis and dysplastic liver cells exhibited strong positive immunostaining for Wnt5a. In contrast, in 65/85 (76.5%) of HCC patients, Wnt5a immunostaining was reduced or absent compared to the levels in adjacent nontumorous (hepatitis and cirrhotic) tissues (Figure 3A). There was a significant negative correlation between Wnt5a expression and tumor stage (P < 0.001), and between Wnt5a and serum AFP (P = 0.016). However, there were no significant associations between Wnt5a protein expression and the other clinicopathological features of HCC patients.

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Immunohistochemical staining for Wnt member 5a (A), β-catenin (B, C) and Ki-67 (D) in hepatocellular carcinoma. Original magnification, 400 ×.

β-catenin protein expression in hepatocellular carcinoma

In non-neoplastic liver tissue, a thin membranous β-catenin signal delineated the hepatocytes, and strong membranous and pale cytoplasmic staining of bile ductules was observed. As shown in Figure 3B and C, altered expressions of β-catenin were found in 68.2% (58/85) of HCC cases. These alterations included reductions in the cellular membrane, increases in the cytoplasm, or both, and nuclear accumulation (in 7%, 6/85). However, no evidence of altered β-catenin expression was found in cirrhotic nodules or dysplastic liver cells in adjacent noncancerous liver tissue. In tumor tissues, altered β-catenin expression was significantly associated with a worsening histopathological tumor grade (P = 0.041) and was not significantly associated with the other clinicopathological parameters.

Correlations among the protein expressions of Wnt5a, Ror2 andβ-catenin

Associations among the protein expression levels of Wnt5a, Ror2 and β-catenin are shown in Table 2. Low cytoplasmic Wnt5a expression was positively associated with low cytoplasmic Ror2 expression (r = 0.411, P < 0.001) and abnormal β-catenin expression (r = 0.254, P = 0.019) in HCC tissue. Similarly, there was a statistically significant correlation between low cytoplasmic Ror2 expression and abnormal β-catenin expression (r = 0.267, P = 0.014).

To investigate the biological functions of proteins in HCC, the Ki-67 LI was assessed in relation to Ror2, Wnt5a and β-catenin status. A strong correlation between a high Ki-67 LI and the reductive loss of Ror2 (r = -0.344, P = 0.002), or Wnt5a (r = -0.278, P = 0.010), but not β-catenin (r = 0.095, P = 0.386) was found (Figure 3D).

A previous study reported that Wnt5a and Ror2 were expressed predominantly in metastatic but not primary lesions of metastatic melanoma, suggesting that Wnt5a and Ror2 might be closely correlated with tumor invasiveness and metastasis[31,34]. To determine whether a similar phenomenon occurs in the metastasis of HCC, three cases of lung metastasis of HCC were included in this study. Immunohistochemical analysis showed that Wnt5a and Ror2 were not expressed in either primary or metastatic lesions (Figure 4A and B), whereas β-catenin- positive staining were detected in the cellular membrane (Figure 4C and D). The Ki-67 LI in tumor tissues was 10%.

Open in New Tab   Full Size Figure   Download Figure

Figure 4 Immunohistochemical staining for Wnt member 5a (A), receptor 2 (B), β-catenin (C, D) in lung metastasis tissues. Original magnification: 400 × in D; 100 × in the others.

The median follow-up was 27.0 mo for survivors (range, 1-54 mo). Three patients were lost to follow-up after surgery and were excluded from the survival analyses. The overall survival curve for the remaining 82 HCC cases is shown in Figure 5A. The estimated 1- and 3-year overall rate of survival was 75% and 44%, respectively. Kaplan-Meier analysis was used to compare the survival rates of HCC patients with tumors expressing low or high levels of Wnt5a and Ror2 and normal or abnormal β-catenin (Figure 5B-D).

Open in New Tab   Full Size Figure   Download Figure

Figure 5 Survival curves of 82 hepatocellular carcinoma patients. A: Overall survival curves of 82 hepatocellular carcinoma (HCC) patients; B: Survival curves of 82 HCC patients with tumors expressing low or high levels of Wnt member 5a (Wnt5a) (log-rank test, P = 0.016); C: Survival curves of 82 HCC patients with tumors expressing low or high levels of receptor 2 (Ror2) (log-rank test, P = 0.007); D: Survival curves of 82 HCC patients with tumors expressing low or high levels of β-catenin (log-rank test, P = 0.045). L: Low expression; H: High expression; N: Normal expression; A: Abnormal expression.

In a univariate Cox proportional hazard regression model analysis (Table 3), tumor stage (P < 0.001), serum AFP (P = 0.036), and the expressions of Wnt5a (P = 0.024) and Ror2 (P = 0.011) were significantly associated with overall survival. Therefore, patients with tumors having a low expression of Wnt5a and Ror2 had a poorer prognosis than those with tumors of high Wnt5a and Ror2 expression.

Multivariate Cox regression analysis (Table 3), the expression levels of Wnt5a (P = 0.020), and β-catenin (P = 0.013) showed a significant association with overall survival. However, a significant correlation between the expression levels of Ror2 and overall survival (P = 0.144), serum AFP and overall survival (P = 0.343) were not demonstrated.

Hepatocellular carcinoma (HCC) is one of the most common cancers in the world. Understanding the molecular biological features of HCC is necessary for early diagnosis and better prognosis. The potential role of Wnt member 5a (Wnt5a) and receptor tyrosine kinase-like orphan receptor 2 (Ror2) in human HCC is receiving increasing attention.

Recent work in a wide of human tumors has indicated that Wnt5a and Ror2 have a critical role in malignant progression. However, little is known about the association of Wnt5a expression with Ror2 and canonical Wnt in HCC. In this study, the authors demonstrate that Wnt5a, in conjunction with Ror2 and β-catenin, may take part in the progression of HCC.

The loss of Wnt5a and Ror2 protein expression in HCC tumor tissue frequently occurs during the progression of HCC and is associated with patient poor prognosis. Wnt5a and Ror2 synergistically execute an anti-tumor effect during the development of HCC. The loss of Wnt5a and Ror2 protein expression was shown to be associated with abnormal β-catenin expression. This is the first study to report an association of Wnt5a expression with Ror2 and β-catenin in HCC.

The study results suggest that protein expression of Wnt5a and Ror2 may be used as clinicopathological biomarkers for prognosis of HCC.

Wnt5a is a non-canonical member of the Wnt family of secreted glycoproteins that acts through the family of frizzled G-protein-coupled receptor, Ror2, to mediate important events during development and cancer.

This paper reported that the loss of Wnt5a and Ror2 protein expression in HCC was associated with poor patient prognosis. Based on reduction in tumors, the authors conclude these markers could be tumor suppressor genes and good prognostic markers for HCC patients. The work is purely descriptive and relevance to clinical practice is significant.