Introduction
Hepatocellular carcinoma (HCC) is the fifth most common malignancy worldwide and causes nearly one million deaths a year (1). Despite the development of several modalities for the treatment of HCC (2–7), including transcatheter arterial embolization, percutaneous ablation, surgical resection, liver transplantation and molecular targeted medicine, the prognosis of patients with HCC still remains relatively poor. One of the major factors responsible for these unsatisfactory outcomes is the high frequency of intrahepatic recurrence after curative treatment (5,8). Intrahepatic recurrence is the result of two mechanisms: intrahepatic metastasis originating from the primary cancer, and a second primary cancer arising through multicentric carcinogenesis. Intrahepatic metastasis may correlate with early recurrence and poor prognosis, whereas multicentric carcinogenesis is associated with a relatively good prognosis (9–11). With the aim of controlling intrahepatic recurrence of HCC, various studies have investigated the molecular mechanisms underlying intrahepatic metastasis (12–17), which frequently occurs at an advanced disease stage, presumably through tumor cell dispersal via the portal vein; there is a strong statistical correlation between the presence of intrahepatic metastasis and the frequency of vascular invasion (18).
Tumor invasiveness may be considered a phenomenon of cell motility. In fact, tumor cell motility plays a central role in carcinoma cell dissemination, and the cytoskeleton, a key structure of the cell machinery, is continuously remodeled during cell movement. In this context, several molecules related to the microtubule (MT)- and actin-dependent dynamics of tumor cells have been investigated as possible predictors of intrahepatic metastasis of HCCs, or targets of preventive therapy. Highly dynamic MTs are distributed randomly throughout the cell periphery, and the less dynamic ones are located between the nucleus and the leading edge of the cell (19). A previous study has revealed that overexpression of HDAC6 increases cell motility, suggesting that deacetylation of at least one cytoplasmic HDAC protein enhances motility (20). It has also been shown that in HDAC6-inhibited cells, MT dynamics are decreased, leading to an increase of focal adhesion accumulation, and thus a decrease in cell motility (21). Moreover, HDAC6 protein can also interact with a different substrate, cortactin, in vivo and in vitro, and both HDAC6 catalytic domains are necessary for the interaction. Cortactin is an acetylated protein found in areas of dynamic actin assembly, such as the leading edge of migrating cells (22). This protein was originally identified as a substrate of Src tyrosine kinase, and plays a role in regulating cell motility. Disruption of HDAC6 leads to hyperacetylation of cortactin and prevents its translocation to the cell periphery, blocks association with F-actin, and impairs cell motility (23). We recently demonstrated that disruption of the HDAC6/NACC1 (nucleus accumbens associated 1) deacetylation system markedly downregulated cell motility through MT and cortactin deacetylation (24). Thus, HDAC6 may act as a mediator between actin- and tubulin-associated proteins to regulate cell motility.
Although overexpression of HDAC6 and its relationship with invasion and metastasis have been documented in several malignancies (24–26), it has not been documented in HCCs. The present study examined the expression of HDAC6 in HCC cultured cells and primary tumors, and investigated its association with migration and invasion activities in vitro and in vivo.
Materials and methods
Cell culture
HCC cell lines were obtained from HSRRB (Health Science Research Resources Bank, Osaka, Japan; HLF, PLC/PRF/5) and IDAC (Institute of Department, Aging and Cancer, Tohoku University, Sendai, Japan; Hep3B). Two human normal hepatocyte cell lines, Hc-cells (Applied Cell Biology Research Institute, Human Hepatocyte Cell culture #3716, Kirkland, WA, USA) and WRL-68 (American Type Culture Collection, Rockville, MD, USA), were obtained commercially and maintained under the recommended conditions.
Surgical specimens and immunohistochemistry
Immunohistochemistry for HDAC6 protein expression was performed on tumor samples from 70 patients with HCC treated between 2006 and 2011 at the Department of Surgery, School of Medicine, Iwate Medical University, Morioka, Japan. The patient characteristics are summarized in Table I. Permission for the study was obtained from the Institutional Review Board (School of Medicine, Iwate Medical University, Morioka, Japan) and written consent was obtained from all patients before surgery.
Surgical specimens were fixed in 10% buffered formalin solution and embedded in paraffin wax, and two or more blocks were made for immunohistochemistry. Sections 4 μm thick were cut, and stained with hematoxylin and eosin. Serial sections were stained with the avidin-biotin system on a Ventana automated immunostainer with the Ventana immunohistochemistry detection system (Ventana Medical Systems, Tucson, AZ, USA), in accordance with the manufacturer’s manual. An anti-HDAC6 antibody (H-300, diluted 1:100, Santa Cruz Biotechnology, Santa Cruz, CA, USA) was used as the primary antibody.
Western blotting
All cell lines were cultured to 70–80% confluence on 10-cm Petri dishes. Cold PBS was added, and the cells were removed from the dishes by scraping. After removal of the supernatants, the cell pellet was dissolved in cell lysis buffer [50 mM Tris-HCl, pH 8.0/150 mM NaCl/1 mM EDTA, pH 8.0/1% Triton X-100/0.1% sodium deoxycholate/0.1% SDS/1 mM PMSF/10 mM NaF/2 mM Na3VO4/1× protease inhibitor complete (Roche Diagnostics GmbH, Mannheim, Germany)]. Cell samples containing equal amounts of protein were mixed with 5× sample buffer, and heated for 5 min at 95°C. Protein was electrophoresed on 4–12% Nu-PAGE for 45 min at 200 V constant voltage, and then transferred onto polyvinylidene difluoride membranes (Invitrogen, Carlsbad, CA, USA) for 1 h at 30 V constant voltage. Membranes were blocked with 5% blocking reagent (Cell Signaling Technology, Danvers, MA, USA) in 1× Tris-buffered saline/Tween-20 buffer for 1 h at room temperature, and immunostained overnight with a primary antibody (HDAC6 H-300, diluted 1:250, Santa Cruz Biotechnology) at 4°C. The membranes were then rinsed with Tris-buffered saline/Tween-20 and incubated with horseradish HRP-conjugated secondary antibodies [anti-rabbit or -mouse IgG (diluted 1:5,000, GE Healthcare, Little Chalfont, UK)] for 1 h at room temperature. Signals were detected with ECL Prime (GE Healthcare) and ChemiDoc XRS (Bio-Rad Laboratories, Hercules, CA, USA). The intensity of the detected signals was measured using 1-D analysis software (Quantity One, Bio-Rad Laboratories). For normalization of the target, glyceraldehyde-3-phosphate dehydrogenase (GAPDH, diluted 1:100; Covance, Princeton, NJ, USA) was used as an internal control.
RNA isolation and reverse transcriptase quantitative PCR
Total RNA was extracted with TRIzol reagent (Invitrogen), and transcribed to cDNA using a SuperScript III first-strand synthesis system (Invitrogen). For quantitative evaluation of relevant mRNAs, we used Custom TaqMan Gene Expression assays (HDAC6, Hs00195869_ m1; Invitrogen), and an ABI PRISM 7500 instrument (Invitrogen). For normalization of the target, GAPDH (Invitrogen) was used as an internal control. All reactions (each containing 3 templates) were run in triplicate, and average fold differences were calculated by normalizing the relative expression (ΔΔCt values) according to ABI User Bulletin #2.
siRNA knockdown of the HDAC6 gene
For silencing of HDAC6 gene mRNA, three predesigned HDAC6-specific siRNA sequences (#1, s19459; #2, s19760; #3, s19461; Silencer Select siRNA, Invitrogen), and control non-specific human siRNAs (Silencer Select Pre-designed siRNA Negative Control #1, 4390843; #2, 4390844, Invitrogen) were used. siRNA transfection was performed using Lipofectamine RNAiMAX Reagent (Invitrogen) in accordance with the manufacturer’s instructions.
Scratch assay
Confluent monolayer cells were scratched to create a wound, and then 0, 24 and 48 h later, three different fields of each wound were photographed with a phase-contrast microscope. Three independent experiments were performed. Measurements of the width of each wound were performed under each experimental condition. At the start of the experiment, the wound size was measured and scored as 100%. After 24 and 48 h, the width of the residual wound was measured and the average percentage of wound closure was calculated by using the free web software ImageJ (http://rsb.info.nih.gov/ij).
Matrigel invasion assay
The cell invasion assay was performed using a BioCoat Matrigel invasion chamber (Becton-Dickinson, Bedford, MA, USA) in accordance with the protocol provided by the manufacturer. After cells in log-phase growth had been incubated with serum-free medium for 12 h, they were detached using trypsin-EDTA. Resuspended cells were added to each chamber at a density of 5×104 cells in 500 μl, and allowed to invade the Matrigel for 24 h at 37°C under a 5% CO2 atmosphere. Cells that had not penetrated the filter were wiped out with cotton swabs, and cells that had migrated to the lower surface of the filter were stained with Quick-Diff stain kit (Symex International Reagents, Co., Ltd., Hyogo, Japan). After two washes with water, the chambers were allowed to air-dry, and the number of invading cells was counted using a light microscope. The degree of invasion was expressed as the average number of migrated cells bound per microscopic field over four fields per assay, and as averages for triplicate experiments.
Statistical analysis
All data are presented as means ± standard error. Correlations between HDAC6 protein expression and clinicopathological data were analyzed by Fisher’s exact test or Kruskal-Wallis test. Mann-Whitney U test for non-parametric samples was used for analyses of biological experimental data. The level of significance was considered to be P<0.05.
Results
Expression of the HDAC6 gene in cell lines and efficiency of knockdown by treatment with siRNA
We first examined HDAC6 mRNA/protein expression in three HCC cell lines and two primary-cultured normal hepatocyte lines (Fig. 1). Under the recommended culture conditions at 60–70% confluency, all of the HCC cell lines exhibited overexpression of HDAC6 mRNA/protein in comparison with normal hepatocytes (Fig. 1). We evaluated the knockdown efficiency of HDAC6-siRNAs (#1, #2 and #3; 10 nM) in one HCC cell line (PCL, Fig. 2). In comparison with negative control siRNA, all the siRNAs caused 80–90% downregulation of HDAC6 expression (Fig. 2).
Migration and invasion activities induced by treatment with HDAC6-siRNA in HCC cell lines
Using #1 HDAC6-siRNA, we then examined the migration activities of the three HCC cell lines by the scratch assay. HDAC6 knockdown significantly decreased tumor cell migration activities of all three lines in comparison with the negative control at 24 and 48 h (P<0.05, Mann-Whitney U-test; Fig. 3). To determine whether HDAC6 knockdown decreased the invasiveness of HCC cells, we performed the Matrigel invasion assay. Treatments with HDAC6-siRNA significantly suppressed the invasiveness of all HCC cell lines in comparison with control siRNA treatment (Fig. 4).
Immunohistochemistry of HDAC6 protein, and relationship between HDAC6 expression and clinicopathological variables in primary HCCs
We immunohistochemically examined HDAC6 protein expression in 70 patients with primary HCCs. Two independent pathologists performed the assessment of immunohistochemical staining. Immunoreactivity for HDAC6 was diffusely positive in both tumor cells and surrounding normal hepatocytes. Overexpression of HDAC6 protein to a level higher than that in the corresponding normal hepatocytes was observed in 14 (20%, Fig. 5C) of the 70 primary HCCs. Lower immunoreactivity for HDAC6 protein was found in 21 (30%, Fig. 5A) of the 70, and the remaining 35 (50%, Fig. 5B) exhibited immunoreactivity equal to that of the corresponding normal hepatocytes. Table II summarizes the relationship between HDAC6 immunoreactivity and clinicopathological variables in all cases. Overexpression of HDAC6 protein was significantly correlated with high clinical stage, number of tumors, vascular invasion, and intrahepatic metastasis (P<0.05) (Table II).
Discussion
Several growth factors and cytokines such as transforming growth factor (27), platelet-derived growth factor (28), epidermal growth factor (29), hepatocyte growth factor (30), and extracellular matrix (31) secreted into the microenvironment surrounding tumor cells are involved in their migration and invasion. These secreted proteins and their related intracellular signaling cascades induce epithelial mesenchymal transition and accelerate the migration/invasion activity of HCC, resulting in intrahepatic metastasis. Moreover, several adhesion molecules and their related proteins such as E-cadherin, ROCK2 and CD24 also are involved (32–34). Apart from proteins secreted into the microenvironment of tumor cells and adhesion molecules, transcriptional factors such as p300 and Snail(35,36) also contribute to the acquisition of metastatic potential by HCC cells. Our present study showed that overexpression of HDAC6, which affects both MT- and actin-dependent cell migration mechanisms, contributed to acceleration of migration/invasion activity in HCC cell lines in vitro and in vivo. In particular, the results of immunostaining of primary HCCs were well correlated with intrahepatic metastasis. HDAC6 was thus suggested to be a newly characterized key player in the control of intrahepatic metastasis.
HDAC6 is a unique protein of the histone deacetylase family, and can affect the function of cytoplasmic non-histone proteins. It is a key regulator of many aspects of cancer biology such as the cell cycle, cell migration, drug resistance and autophagy, thereby making HDAC6 an attractive target for cancer therapy (37). Although HDAC6 protein is expressed in the liver as well as the heart, kidney, testis, brain, and pancreas (38), there has been little information about the significance of HDAC6 in HCC tumorigenesis. It is well known that HDAC inhibitors, such as TSA and SAHA, block invasive cell motility, and therefore it is anticipated that they might be applicable for control of intrahepatic metastasis. However, most of these molecules act by altering gene expression via hyperacetylated HDAC nuclear substrates, such as histones or transcription factors, and the spectrum of targeted molecules is broad. Therefore, development of inhibitors that are more selective in targeting intrahepatic metastasis is warranted. An HDAC6 inhibitor known as tubacin (tubulin acetylation inducer) was isolated through a multidimensional chemical genetic screen of 7,392 small molecules and a cell-based assay targeting the acetylation activity of proteins other than histones (39,40). Unlike other histone deacetylase inhibitors, tubacin was found to inhibit the deacetylation of MT in mammalian cells without affecting the level of histone acetylation, gene expression, or cell cycle progression (13,14). Furthermore, using scratch assay and trans-Matrigel migration assays, another group has demonstrated that NK84-mediated inhibition of HDAC6 in ovarian cancer cell lines retarded cell spreading and inhibited cell migration, respectively (41). Recently, the effectiveness of combination therapy using an HDAC6-selective inhibitor (ACY-1215) and bortezomib has been demonstrated in a preclinical trial (42). Thus, a new class of agents targeting HDAC6 is currently being developed, and their efficacy is being tested.
The functions of HDAC6 in cell migration/invasion activity may depend on decacetylation activity targeting α-tubulin and cortactin. Both proteins are also deacetylated by SIRT2, which belongs to another class of the histone deacetylase family. Using SIRT2-specific siRNA combined with tubacin treatment, Zuo et al(25) have demonstrated that cell migratory and invasive abilities can be dramatically suppressed. Moreover, SIRT2-deficient mice show gender-specific tumorigenesis, females primarily developing mammary tumors, and males developing more HCCs (43). The significance of SIRT2 should therefore be examined in human HCC tumorigenesis, including its relationship with intrahepatic metastasis.