Increased Expression of Wiskott-Aldrich Syndrome Protein Family Verprolin-Homologous Protein 2 Correlated with Poor Prognosis of Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) ranks the sixth among the most common malignancies worldwide (626,000 or 5.7% of new cancer cases; ref. 1) and is the second leading cause of cancer death among males in China (2). It is also one of the most deadly human carcinomas due to its high recurrence and metastasis rate (3–8). The recurrence and metastasis of HCC is a multistep process that often involves many complex biological and pathologic events (9). Despite extensive clinical as well as basic research efforts, the mechanisms for HCC metastasis remain not well understood.

Available information suggests that tumor metastasis is to a large extent attributable to the ability of cell migration (10, 11), a well-documented cellular process dependent on the dynamics of actin filaments formed through actin polymerization (12). During the formation of actin filaments, Arp2/3 complex acts downstream of the Wiskott-Aldrich syndrome protein (WASP) family, which consists of WASP subfamily (WASP and N-WASP) and WASP family verprolin-homologous protein (WAVE) subfamily (WAVE1, WAVE2, and WAVE3; refs. 13–15). WAVEs are thought to act downstream of the Rac GTPase, connecting Rac activation to the induction of Arp2/3-mediated actin polymerization (16). Coupling of Rac to WAVE effectors involves interactions of these effectors with Abi1 (17, 18). In addition, phosphatidylinositol-3,4,5-trisphosphate also recruits WAVE2 to the polarized membrane (19). Rac appears to be the upstream protein responsible for the regulation of WAVE-dependent actin filament formation (20). The Rac signaling pathway has been reported to be important to HCC cell motility (21). Interestingly, a recent study using murine melanoma cells B16F10 suggested a critical role for the Rac-WAVE2 signaling in the invasive and metastatic phenotypes (22). Being essential for the actin polymerization at the leading edge of cell membrane, WAVE2 was also shown to be indispensable for driving cell migration (23). Although these previous studies have suggested a role for WAVE2 in tumor cell migration, its relevance to human HCC remains unknown.

Our early study using cDNA microarray to profile gene expression pattern in HCC suggested an elevated expression of WAVE1 in HCC tissues over in paracarcinomatous liver tissue (PCLT; ref. 24; data not shown). In view of the similar function of WAVE family proteins in cell migration and the essential coordination of WAVE1 and WAVE2 activities for formation of proper actin structures in stable lamellipodia (25), we carried out the present study to determine the expression patterns of WAVE2 in HCC tissues and to explore the relationship of WAVE2 expression with the clinicopathologic features and prognosis of HCC. Our previous studies have shown that HCC can be divided into solitary large HCC (SLHCC; diameter >5 cm, only one nodule, often grows expansively within an intact capsule or pseudocapsule, and has a relatively benign biological behavior) and nodular HCC (NHCC, tumor nodule number ≥2; ref. 26); there was a significant difference in the incidence of metastasis between them: SLHCC is characterized with a low incidence of metastasis, whereas NHCC is much more metastatic and often associated with a very poor prognosis. With this difference, we also compared the WAVE2 expression between NHCC and SLHCC to further analyze the correlation between WAVE2 and the metastasis potential of HCC.

Patients and specimens. In the present study, specimens of HCC tissues were obtained from 112 HCC patients who underwent hepatectomy at the Department of Surgery, Xiangya Hospital of Central South University (Changsha, People's Republic of China) from February 2000 to September 2003. The 112 patients included 104 males and 8 females with a median age of 46 years (range, 18-73 years). Among these 112 cases of HCCs, matched HCC and PCLT specimens from 31cases were also collected and immediately frozen in liquid nitrogen, which were subsequently stored at −80°C for reverse transcription-PCR (RT-PCR) and Western blotting analysis. Six cases of normal liver samples were obtained from patients with liver cavernous hemoangioma as controls. All specimens were paraffin embedded and stained by H&E. The diagnoses of all these patients were confirmed by histopathologic examination. The clinicopathologic variables, such as tumor diameter, the number of tumor nodules, tumor capsule, histopathologic classification, and vein invasion, were recorded. Using these variables, 31 cases from which the paired samples were collected were further divided into two groups according to the criteria described in Introduction: SLHCC (n = 11) and NHCC (n = 20). Prior informed consent was obtained from the patients for collection of liver specimens in accordance with the guidelines of Xiangya hospital, and the study protocols have been approved by the Ethics Committee of the Central South University.

Follow-up. Follow-up data were obtained after hepatic resection for all 112 patients. The follow-up period was defined as the interval between the date of operation and that of patient's death or the last follow-up. Deaths from other causes were treated as censored cases; the follow-up time ranged from 30 to 1130 days, with a median follow-up time of 380 days.

RT-PCR. Total RNA was isolated as described previously (27) by using Trizol reagent (Life Technologies, Grand Island, NY). Total RNA (2 μg) was reverse transcribed by Moloney murine leukemia virus reverse transcriptase (Fermentas, Burlington, Ontario, Canada). Reverse transcriptase product (2 μL) was amplified by PCR using the following conditions: 95°C for 5 minutes and then 32 cycles of 95°C for 40 seconds, 55°C for 30 seconds, 72°C for 55 seconds, extension 72°C for 5 minutes. WAVE2 primers: 5′-AGCCTTCAGAAGTTCCACCA-3′ (forward) and 5′-CTGCAGCATCTTCTCCTTCC-3′ (reverse). The expression of a constitutive glyceraldehyde-3-phosphate dehydrogenase gene was determined as a control using the following primers: 5′-CTGCAGCATCTTCTCCTTCC-3′ (forward) and 5′-CAAAGTTGTCATGGATGACC-3′ (reverse). PCR product (5 μL) was then electrophoresed on 1.5% agarose gel, and the intensity of bands was quantified by Eagle Eye II laser densitometry program (Strategene, La Jolla, CA). WAVE2 gene expression was presented by the relative intensity of the PCR product bands from target sequences to that from the glyceraldehyde-3-phosphate dehydrogenase gene. PCR experiments were done in triplicate.

Western blotting. The total protein was extracted from HCC or PCLT and the concentration of protein was determined by using bicinchoninic acid protein assay kit (Pierce, Rockford, IL). Total protein (100 μg) was separated by SDS-PAGE and then transferred onto nitrocellulose membrane (Millipore, Bedford, MA). The blotted membranes were incubated with goat anti-human WAVE2 polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA) diluted at 1: 500. After washing, the membranes were incubated with a 1: 3,000 dilution of horseradish peroxidase–linked mouse anti-goat antibody (KPL, Gaithersburg, MD). The blots were developed by enhanced SuperSignal West Pico chemiluminescence (Pierce). β-Actin was also determined by using the specific antibody (Sigma, St Louis, MO) as a loading control. All experiments were carried out in triplicate and the levels of WAVE2 protein were quantified by densitometry (Strategene).

Immunohistochemistry. Tissue sections of 4-μm thick were cut and baked at 60°C for overnight, deparaffinized in xylene, and rehydrated through graded ethanol. Next, 3% hydrogen peroxide was applied to block the endogenous peroxidases for 30 minutes and sections were subjected to microwave heat–induced antigen retrieval in EDTA (1 mmol/L; pH 8.0) at high power twice, each for 7 minutes. After rinsing with PBS, the sections were incubated with normal goat serum for 30 minutes at 37°C to block nonspecific binding. The samples were then incubated at 37°C for 30 minutes with goat anti-human WAVE2 polyclonal antibody (1: 100 dilution) and the second antibody for 30 minutes at 37°C. The streptavidin-biotin-peroxidase complex tertiary system (Boster, Wuhan, People's Republic of China) was used according to the manufacturer's instruction. All slides were visualized by applying 3,3-diaminobenzidine tetrahydrochloride for 2 minutes and then counterstained with hematoxylin. Negative control slides were probed with normal goat serum under the same experimental conditions. The immunohistochemistry results were verified by Western analysis of same set of human specimens. A very nice correlation of WAVE signals between immunohistochemistry and Western analyses was obtained (data not shown). The slides were examined independently by two pathologists. The intensity of WAVE2 protein staining in HCC specimens was classified using a four-point scale: 0, ≤10% positive cells; 1+, 11% to 25% positive cells; 2+, 26% to 50% positive cells; and 3+, ≥51% positive cells; the expression levels of the WAVE2 protein were thus divided into low expression (0 or 1+) and high expression (2+ or 3+).

Statistical methods. Quantitative values were presented as mean ± SD or median (range). Independent Student's t test was used to compare WAVE2 mRNA and protein expression in HCC and PCLT samples. A similar comparison was made between SLHCC and NHCC tissues. Spearman's correlation coefficient was used to examine the correlations between expression levels of WAVE2 mRNA and protein as well as the relationships between WAVE2 mRNA and protein expression levels and clinicopathologic variables of 31 cases of HCCs. The correlations between WAVE2 immunohistochemical staining and clinicopathologic variables of 112 cases of HCCs were analyzed by Mann-Whitney U test. The relation of WAVE2 expression to the overall survival of HCC patients was assessed by Kaplan-Meier estimates, and high and low expression groups were compared by the log-rank test. The Cox proportional hazard model was done to identify factors that were independently associated with survival. In this model, a stepwise selection was used for variable selection with entry and removal limits of P < 0.05 and P > 0.10, respectively. All statistical evaluations were done using the SPSS (version 10.0; Chicago, IL). All tests were two tailed and P < 0.05 was considered statistically significant.

Elevated expression levels of WAVE2 mRNA and protein in HCC tissues. RT-PCR analysis using the WAVE2-specific primers indicated that WAVE2 mRNA was readily detectable in all HCC and PCLT tissues of 31 cases. However, HCC tissues expressed significantly higher levels of WAVE2 than PCLT tissues (1.32 ± 0.50 versus 0.44 ± 0.13; P < 0.001; Fig. 1A). Consistent with the mRNA expression, the WAVE2 protein was also detected in all 31 samples derived from HCC and PCLT tissues, and the average relative expression in HCC tissues was significantly higher than that in the corresponding PCLT tissues (0.90 ± 0.22 versus 0.42 ± 0.12; P < 0.001; Fig. 1A). When compared with normal liver tissues, the levels of WAVE2 mRNA and protein in PCLT tissues were slightly elevated but not statistically significant (0.44 ± 0.13 versus 0.41 ± 0.05; P > 0.05). Significantly, we found that the expression levels of both WAVE2 mRNA and protein were in general higher in NHCC than in SLHCC [for mRNA, 1.51 ± 0.51 versus 0.97 ± 0.21 (P = 0.001); for protein, 1.02 ± 0.18 versus 0.70 ± 0.13 (P < 0.001); Fig. 1B]. Furthermore, there was a significantly positive correlation between expression levels of WAVE2 mRNA and protein in HCC by Spearman's correlation coefficient analysis (r = 0.788; P < 0.001; Fig. 2).

Correlations between WAVE2 expression and clinicopathologic variables. Among 112 HCC samples analyzed immunohistochemically, there were 30 (26.8%) cases with >50% positive cells (scores 3+), 41 (36.6%) cases with 26% to 50% positive cells (scores 2+), 25 (22.3%) cases with 11% to 25% positive cells (scores 1+), and 16 (14.3%) cases with 0% to 10% positive cells (scores 0). Figure 3 shows the representative immunostaining of the specimens.

We used the Mann-Whitney U test to assess the correlations between the staining intensity of WAVE2 protein and clinicopathologic variables of HCC. Interestingly, the WAVE2 protein staining intensity in HCC with multiple tumor nodules were significantly higher than those with solitary tumor nodule (P = 0.008). HCC without capsular also showed significantly stronger WAVE2 staining than those with intact capsular (P = 0.035) Furthermore, the increased staining intensity significantly correlates with the Edmondson-Steiner grade of HCC (P = 0.009). A difference in expression of the WAVE2 protein between the HCCs with vein invasion and those without vein invasion was also evident (P = 0.023). However, the WAVE2 protein staining intensity showed no significant relationship with gender, age, liver cirrhosis, and tumor size (P > 0.05; Table 1). In 31 HCC cases whose WAVE2 mRNA and protein expression levels were determined, there was a positive correlation with the immunohistochemistry results. The WAVE2 expression also significantly correlated to tumor nodule number, capsular formation, Edmondson-Steiner grade, and vein invasion (P < 0.05; Table 2).

Relationship between WAVE2 expression level and prognosis. Based on the immunohistochemistry data, we divided 112 cases of HCCs into the low expression group (immunohistochemistry score 0 and 1+; n = 41) and the high expression group (immunohistochemistry scores 2+ and 3+; n = 71). The Kaplan-Meier method was used to analyze the correlation of WAVE2 expression level and the prognosis of HCC patients. Our results indicated that the higher level of WAVE2 expression correlated with a shorter survival time (326 versus 512 days, median survival time) and the overall survival rate for the patients with low and high WAVE2 level was significantly different (P = 0.003, log-rank test; Fig. 4).

To do an independent test of the association between WAVE2 expression and overall survival, univariable and multivariable Cox regression analyses were used to identify factors that may predict survival after hepatectomy. The results indicate that high levels of WAVE2 expression [relative risk (RR), 1.96; P = 0.004], vein invasion (RR, 2.6; P = 0.009), tumor nodule number (RR, 1.65; P = 0.019), high serum α-fetoprotein level (RR, 1.63; P = 0.031), absence of capsule (RR, 1.72; P = 0.033), and high Edmondson-Steiner grade (RR, 1.72; P = 0.037) were all significantly associated with survival. Gender, age, hepatitis B virus association, serum alanine aminotransferase level, cirrhosis of the liver, Child-Pugh classification, type of hepatectomy done, margin status, and tumor size were not significantly correlated with survival (Table 3). By multivariable Cox regression analysis, high WAVE2 expression (RR, 1.68; P = 0.038), high serum α-fetoprotein level (RR, 1.63; P = 0.041), multiple tumor nodules (RR, 1.59; P = 0.045), and vein invasion (RR, 1.55; P = 0.047) were found to be independent prognostic factors for survival. The other clinicopathologic variables did not add any independent prognostic information.

The result of our previous gene array study that the expression of WAVE1 was higher in human HCC tissues over PCLTs (24) led us to ask whether the expression of WAVE2, as an important member of the WAVE family, might also increase in HCC. To test this, we examined the WAVE2 expression levels of mRNA and protein in 31 paired HCC and PCLT samples. The results obtained from RT-PCR and Western blotting analyses revealed significantly elevated expression levels of WAVE2 mRNA and protein in HCC tissues when compared with PCLTs. Although we also found a slightly increased expression level of WAVE2 mRNA as well as protein in PCLT tissues over that in normal liver tissues, the difference was not statistically significant. Together, the results indicated a significantly increased expression of WAVE2 in HCC.

Our previous studies have shown that, according to distinct phenotypes, HCC can be divided into two categories, NHCC and SLHCC. NHCC displays significantly greater potency in invasion and metastasis than SLHCC does (24, 26, 28). In the present study, we also examined the expression patterns of WAVE2 in SLHCC and NHCC to explore the association of WAVE2 expression with the metastasis of HCC. A generally higher level of WAVE2 expression in NHCC than in SLHCC was observed, which correlates with the higher metastasis potential of NHCC than that of SLHCC, supporting the correlation between WAVE2 expression levels and metastasis in HCC. Our finding agrees with the results of a mouse study that WAVE2 was essential for invasion and metastasis of mouse melanoma (22). Having used Western blot to verify the immunohistochemistry method, we further investigated the relationship of WAVE2 expression with clinicopathologic features by immunohistochemistry analysis of 112 HCC cases. Among the variables that included gender, age, liver cirrhosis, tumor size, tumor nodule number, capsular formation, Edmondson-Steiner grade, and vein invasion, the expression of WAVE2 seemed to correlate with the multiple tumor nodules, the absence of capsular formation, vein invasion, and high Edmondson-Steiner grade of HCCs. The number of tumor nodules is thought to contribute to recurrence of patients with HCC (29). Capsulation formation of HCC reflects reduced invasiveness and improved survival (30). The cell differentiation status (the Edmondson-Steiner grade) and vein invasion are highly correlated with the invasion and metastasis as well as prognosis of HCC (2, 5, 31, 32).

These findings support a correlation of WAVE2 expression with the metastasis of HCC, suggesting a prognostic implication of WAVE2 expression in HCC. With our data implicating WAVE2 expression as a potential prognostic marker for HCC patients, we used the Kaplan-Meier method to assess the prediction power of WAVE2 expression level for prognosis of patients with HCC. Based on the results of immunohistochemical staining, we divided the total 112 cases of HCC into the high WAVE2 expression and the low WAVE2 expression group. The HCC patients within the low WAVE2 expression group in general had a better prognosis than those within the high WAVE2 expression group. To test whether the WAVE2 expression level could serve as an independent prognostic factor for HCC, we fitted the WAVE2 expression level as well as the other 14 conventional clinicopathologic variables (33, 34) into the Cox regression model. Multivariable Cox regression analysis indicates that the WAVE2 expression, serum α-fetoprotein level, number of tumor nodules, and vein invasion are independent factors for prediction of HCC survival.

In summary, this is the first report to show that elevated WAVE2 expression levels correlate to the poor prognosis of human HCC, which suggests that WAVE2 may serve as a prognosis marker for HCC. Further studies are required to determine how WAVE2 expression is up-regulated in HCC.

We thank Weiqun Lu for collecting patient survival data.