Rho-kinase 2 is frequently overexpressed in hepatocellular carcinoma and involved in tumor invasion


  • Carmen Chak-Lui Wong,

    1. Department of Pathology, Liver Cancer and Hepatitis Research Laboratory and SH Ho Foundation Research Laboratory, The University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Chun-Ming Wong,

    1. Department of Pathology, Liver Cancer and Hepatitis Research Laboratory and SH Ho Foundation Research Laboratory, The University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Edmund Kwok-Kwan Tung,

    1. Department of Pathology, Liver Cancer and Hepatitis Research Laboratory and SH Ho Foundation Research Laboratory, The University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Kwan Man,

    1. Department of Surgery, The University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Irene Oi-Lin Ng

    Corresponding author
    1. Department of Pathology, Liver Cancer and Hepatitis Research Laboratory and SH Ho Foundation Research Laboratory, The University of Hong Kong, Hong Kong
    • Department of Pathology, University Pathology Building, Room 127B, The University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong
    Search for more papers by this author
    • fax: (852)-2872-5197.

    • I. O. L. N. is Loke Yew Professor in Pathology.

  • Potential conflict of interest: Nothing to report.


Deregulation of Rho family small guanosine triphosphatases has been implicated in human carcinogenesis. Rho-kinases are downstream effectors of Rho guanosine triphosphatases in the regulation of cytoskeletal reorganization and cell motility. However, their functions in human cancers remain elusive. In this study, we aimed to investigate the role of Rho-kinases in hepatocellular carcinoma (HCC) tumor progression and invasion. We first examined the expression of the two Rho-kinases (ROCK1 and ROCK2) in human HCC, and found that ROCK2 was frequently overexpressed in primary HCCs (22/41 [53.66%]). Clinico-pathological analysis revealed that overexpression of ROCK2 was significantly associated with the presence of tumor microsatellite formation (P = 0.005), suggesting that deregulation of ROCK2 may contribute to the intrahepatic metastasis of HCC. Consistently, we demonstrated that stable overexpression of ROCK2 significantly enhanced cell motility and invasiveness in HCC cells. Conversely, stable knockdown of ROCK2 by short hairpin RNA approach remarkably reduced HCC cell migration and invasion. Moreover, orthotopic liver xenograft models provided further support that stable knockdown of ROCK2 suppressed HCC invasion in vivo. Stable knockdown of ROCK2 in HCC cells significantly inhibited Golgi reorientation, myosin phosphatase phosphorylation, and formations of stress fibers, filopodia, and lamellipodia; these molecular and cellular events are crucial for cell motility and cancer invasion. Conclusion: Our results indicate that ROCK2 was overexpressed in human HCCs, and this overexpression was associated with a more aggressive biological behavior. Our findings also demonstrate that ROCK2 played a significant role in regulating cytoskeletal events and contributed to the invasion of HCC. (HEPATOLOGY 2009.)

Hepatocellular carcinoma (HCC) is a prevalent malignancy worldwide and causes about 500,000 deaths per annum. Cancer metastases, both intrahepatic and extrahepatic, are major factors in the mortality of HCC patients. Nonetheless, the molecular mechanisms underlying HCC metastasis remain largely unclear. The Rho-kinases (ROCKs) are immediate downstream effectors of Rho, a member of the Rho family small guanosine triphosphatases. ROCKs play important roles in many physiological functions, including neurite growth retardation,1, 2 prenatal and postnatal development,3–5 and smooth muscle contractions.6 ROCK activity is regulated by distinct members of Rho guanosine triphosphatases (Rho), and the Rho/ROCK pathway also participates in regulating cytoskeletal signaling events and is crucial to cell motility.7 Deregulation of Rho/ROCK pathway has been recently implicated in tumor progression and, particularly, tumor metastasis. Among all the Rho family members, RhoA is the best characterized member in cancers and positively regulates ROCK activity. RhoA was found to be significantly overexpressed in various cancers, including HCC8 and breast,9 lung,10 colon,10 head and neck,11 testicular,12 and bladder cancer.13 The ROCK family was also shown to be involved in various cancer types. Overexpression of the two members of the ROCK family (ROCK1 and ROCK2) has been demonstrated in testicular and bladder cancers at the protein level.12, 13 In addition, it has been shown that ROCK inhibitor (Y27632) treatment is sufficient to abolish chemotactic migration in pancreatic cancer cell lines.14 Likewise, blockade of ROCK activity via Y27632 or overexpression of a dominant negative mutant of ROCK1 has been shown to suppress actomyosin activity, peritoneal invasion, and intrahepatic metastasis in rat and SCID mouse models.15–17 Although up-regulation of Rho/ROCK is often associated with aggressive tumor behavior,18, 19 the exact roles of Rho-kinases in HCC have not been fully elucidated.

In this study, we investigated the functional implications and molecular mechanisms of ROCK2 in HCC development and invasion. We found that ROCK2 was frequently overexpressed in human HCCs and was closely associated with tumor microsatellite formation, a pathologic feature of intrahepatic metastasis. We also demonstrated that ROCK2 promoted HCC migration and invasion both in vitro and in vivo. We further showed that ROCK2 contributed to HCC invasion by regulating events associated with cell directional movement indicated by Golgi reorientation, and actin polymerization such as formations of filopodia, lamellipodia, and stress fibers. These results suggested an important role of ROCK2 in HCC invasion and the potential of ROCK2 as a chemotherapeutic target for treatment of HCC.


GFP, green fluorescent protein; HCC, hepatocellular carcinoma; MMP, matrix metalloproteinase; MYPT1, myosin phosphatase target subunit; PCR, polymerase chain reaction; ROCK1, Rho-kinase 1; ROCK2, Rho-kinase 2.

Materials and Methods

Cell Lines and Patient Samples.

Human HCC cell lines BEL7402 and MHCC97L were gifts from Shanghai Institute of Cell Biology, Chinese Academy of Sciences, and Fudan University (Dr. Z. Y. Tang) of Shanghai, respectively. Human HCC and their corresponding nontumorous liver samples were collected at the time of surgical resection at Queen Mary Hospital, the University of Hong Kong, from 1991 to 2000. All samples, after collection from surgical resection, were snap-frozen in liquid nitrogen before storage at −80°C.


Anti-ROCK2 antibody used for protein expression analysis in human samples was purchased from Anaspec Inc. (San Jose, CA) and anti-ROCK2 antibody for protein expression analysis in cell lines was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-MYPT antibody and anti–phospho-MYPT (Thr853) antibody were obtained from Upstate (Lake Placid, NY), anti–β-actin antibody was obtained from Cell Signaling (Danvers, MA), and anti–green fluorescent protein (GFP) antibody and anti-ROCK1 antibody were obtained from Santa Cruz.

Protein Extraction and Western Blotting.

For western blotting, cells were lysed with sodium dodecyl sulfate buffer (for detection of phospho-MYPT, MYPT) or NP40 in NET buffer (for detection of ROCK1 and ROCK2). Proteins from human samples were extracted with RIPA buffer. Proteins were resolved via sodium dodecyl sulfate–polyacrylamide gel electrophoresis and blotted onto a nitrocellulose membrane. The membrane was incubated with the indicated primary antibody, followed by incubation with anti-mouse immunoglobulin G or anti-rabbit immunoglobulin G (GE Healthcare, Buckinghamshire, UK). Protein expression was detected with the ECL detection system (GE Healthcare) according to the manufacturer's protocol. Images were scanned and intensity of bands was quantified with AlphaEaseFC software.

Clinico-pathological Correlation and Statistical Analysis.

The clinico-pathological features of HCC patients (tumor size, cellular differentiation according to the Edmondson grading, venous invasion without differentiation into portal or hepatic venules, direct liver invasion, tumor microsatellite formation, tumor encapsulation, number of tumor nodules, and serum hepatitis B surface antigen status) were analyzed using SPSS for Windows 14.4 (SPSS, Inc., Chicago, IL) as described.20 Categorical data were analyzed via Fisher's exact test, whereas an independent t test was used for continuous data. Test results were considered significant at P < 0.05.

Plasmid Construction.

Constitutively active ROCK2 (1-581 aa) was amplified via polymerase chain reaction (PCR) from HepG2 complementary DNA (forward primer, 5′-CCG CTC GAG TAT GTT ATC CCA ATG CCA CTG A-3′; reverse primer, 5′-AAC TGC AGT ATC AGA CTC TGT T- 3′). Purified PCR product was then cloned into pEGFP vector via PstI and XhoI digestions. Vector-based short hairpin RNA specifically targeting ROCK2 was generated by PCR-based strategy based on the sequence of small interfering RNA purchased from Dharmacon (Chicago, IL). In brief, U6 promoter sequence served as template in which the short hairpin RNA sequence was included in the reverse primer (forward primer: 5′-GAT TTA GGT GAC ACT ATT AG-3′ and reverse primer: 5′-AAA AAA GTT AGT GTC CTA TTA GTT TCC AAG CTT CGA AAC TAA TAG GAC ACT AAC GGT GTT TCG TCC TTT CCA CAA-3′). PCR product was cloned into pCR4-TOPO (Invitrogen, Carlsbad, CA) and subsequently subcloned into pDSRed vector via AseI digestion.

Establishment of ROCK2 Stably Expressing Cells and ROCK2 Stable Knockdown Cells.

One day before transfection, 2 ×105 cells were seeded onto 35-mm plates. Plasmids were transfected into cells with FuGene 6 reagent (Roche, Basel, Switzerland). After 24 hours, transfected cells were spread onto 100-mm culture dish at 1:100 dilution. To select for stable transfectants, cells were cultured in Dulbecco's modified Eagle's medium high-glucose medium with 0.7 mg/mL G418 for 4 weeks. Clones with G418 resistance and fluorescent label (DSRed or GFP) were selected and expanded. Stable overexpression and knockdown of ROCK2 were confirmed via western blotting with anti-GFP and anti-ROCK2 antibodies, respectively.

Cell Migration and Invasion Assay.

Transwell assay and wound healing assay were performed as described.21 Cell invasion assay was performed with precoated cell invasion kit (Chemicon International, Temecula, CA), or self-coated Matrigel (BD Biosciences, Sparks, MD) on the upper surface of the transwell chamber. The invasive cells that had invaded through the extracellular matrix layer to the lower surface of the membrane were fixed with methanol and stained with crystal violet. Photographs of three randomly selected fields of the fixed cells were captured, and cells were counted. Experiments were repeated independently three times.

Immunofluorescence Microscopy and Scanning Electron Microscopy.

Cells were seeded onto coverslips and incubated overnight at 37°C in a CO2 incubator. For immunofluorescence study, cells were fixed with 4% paraformaldehyde in phosphate-buffered saline and permeabilized with 0.2% Triton X-100 in phosphate-buffered saline. Fixed cells were incubated with 1:2,000 fluorescein isothiocyanate–conjugated phalloidin (Sigma, St. Louis, MO) or antibodies as indicated. Cells were counterstained with 4',6-diamidino-2-phenylindole (Calbiochem, San Diego, CA) and mounted with Vectashield antifade mountant (Vector Laboratories, Burlingame, CA). Images were captured under 1,000 magnification by a fluorescence microscope connected to a charge-coupled device camera (Leica, Wetzlar, Germany). For scanning electron microscopy, cells were fixed with 1% osmium tetroxide and 2.5% glutaldehyde followed by stepwise ethanol dehydration. After the step of critical point dry, slides were mounted on silver paste. Images were scanned and captured under ×5,000 and ×8,000 magnification using a Stereoscan 440 scanning electron microscope (Oxford Instruments, Cambridge, UK).

Orthotopic Liver Implantation in Nude Mice.

To examine the metastatic potential of ROCK2 stable knockdown clones, 1 × 106 cells (BEL7402) and 4 × 106 (MHCC97L) in 0.1 mL of phosphate-buffered saline, respectively, were injected subcutaneously into the flanks of BALB/C nude mice. After 4 weeks, the subcutaneous tumors were resected and diced into 1 mm3 cubes, which were then implanted in the left lobes of the livers of the nude mice. For xenografts derived from BEL7402 and MHCC97L cells, the animals were sacrificed and examined 12 and 6 weeks, respectively, after implantation. For in vivo tracking, the MHCC97L cells were stably transfected with firefly luciferase. Tumor metastasis of MHCC97L xenografts was detected weekly with bioluminescent signaling during the course of study. One hundred milligrams per kilogram D-luciferin (Xenogen, Hopkinton, MA) was injected into the peritoneal cavities of the mice, and bioluminescence was detected by IVIS 100 Imaging System (Xenogen, Hopkinton, MA). For ex vivo organ imaging, mice were injected with 100 mg/kg D-luciferin 5 minutes intraperitoneally before necropsy, and the excised organs were imaged. These experiments were performed according to the Animals (Control of Experiments) Ordinance (Hong Kong) and the Institute's guidance on animal experimentation.

Histo-pathological Analysis.

Livers and lungs of the sacrificed mice were harvested and fixed in 10% formalin followed by 75% ethanol before paraffin embedding. Five-micrometer-thick paraffin sections were cut and stained with hematoxylin-eosin for histological examination.

Golgi Reorientation Assay.

Golgi reorientation assay was performed as described elsewhere.22 To record the position of the Golgi apparatus in migrating wound edge cells, wounds were created to confluent monolayers of BEL7402 cells on coverslips. Cells were then allowed to migrate for 5 hours. The Golgi apparatus and nucleus were stained with 10 μg/mL lectin HPA Alexa Fluor 488 conjugates (Invitrogen) and 4',6-diamidino-2-phenylindole (Calbiochem, San Diego, CA), respectively. Wound edge cells were scored positive if the Golgi apparatus located in front of the nucleus in the direction toward the wound. Experiments were repeated independently three times. A range of 353 to 737 cells were counted in total for each wound column.


ROCK2 Is Overexpressed in Human HCC and Is Associated with Aggressive Tumor Behavior.

We examined ROCK1 and ROCK2 protein expression levels in 41 pairs of human HCC and their corresponding nontumorous livers via western blotting. ROCK2 was found to be overexpressed in 53.66% of the HCC cases (Fig. 1). In contrast, there was no significant difference in ROCK1 expression levels between HCCs and their nontumorous livers (Supplementary Fig. 1). Upon clinico-pathological correlation analysis, overexpression of ROCK2 was found to be closely and positively associated with the presence of tumor microsatellite formation (P = 0.005, Fisher's exact test), which is a feature of intrahepatic metastasis (Table 1). These findings indicate that overexpression of ROCK2 is frequent in human HCCs and might be related to tumor metastasis.

Figure 1.

ROCK2 is overexpressed in human HCC. Western blot analysis of representative HCC cases with ROCK2 overexpression. β-Actin was used as a loading control. NT, nontumorous liver; T, tumor.

Table 1. Clinico-pathological Correlation of ROCK2 Overexpression in Human HCC
 Without Overexpression (No. of Cases)With Overexpression (No. of Cases)P Value
  • *

    P < .05.

Tumor size   
 ≤5 cm620.115
 >5 cm1320 
No. of tumor nodules   
Tumor encapsulation   
Cellular differentiation by Edmondson grading   
Venous invasion   
Tumor microsatellite formation   
Direct liver invasion   
Hepatitis B surface antigen   

ROCK2 Enhances Cell Migration and Invasion of HCC Cells in Vitro.

To assess the effect of ROCK2 on HCC cell migration, constitutively active ROCK2 was stably transfected into the BEL7402 HCC cell line, and the expression was confirmed via western blotting (Fig. 2A, Supplementary Fig. 2). We found that ROCK2 stably expressing cells (GFP-ROCK2C11) displayed a significant increase in cell migration ability compared with the vector control (P < 0.001, t test) (Fig. 2B). In addition, using invasion chamber coated with a thin layer of extracellular matrix, we further demonstrated that overexpression of ROCK2 substantially enhanced the invasiveness of HCC cells, as indicated by a marked increase in the number of invaded cells (P < 0.001, t test) (Fig. 2C). These findings were consistent with the correlation we observed between overexpression of ROCK2 and the pathological feature of intrahepatic metastasis in HCC patients and suggested that overexpression of ROCK2 positively regulated HCC cell motility and invasiveness.

Figure 2.

Overexpression of ROCK2 enhances HCC cell migration and invasion. (A) Ectopic overexpression of GFP-ROCK2 fusion protein in BEL7402 stable clones. GFP (lower left arrow) and GFP-ROCK2 fusion (upper right arrow) expression were confirmed via western blot analysis with GFP antibody (see also Supplementary Fig. 2 for full gel picture). (B) Transwell cell migration assay. 5 × 104 cells were seeded onto the upper chamber of the Transwell and allowed to migrate for 8 hours. The number of migrated cells was significantly increased in ROCK2 stable transfectant (GFP ROCK2C11) as compared with vector control (GFP) (P < 0.001, t test). (C) Invasion assay. 1 × 105 cells were seeded and allowed to invade through the extracellular matrix precoated invasion chamber kit for 24 hours. The number of invaded cells was significantly increased in ROCK2 stably overexpressing cells (GFP ROCK2C11) as compared with vector control (GFP) (P < 0.001, t test). Error bars indicate the standard error of the mean.

To further substantiate the notion that ROCK2 was implicated in HCC cell migration and invasion, we stably knocked down ROCK2 expression in BEL7402 by short hairpin RNA (Fig. 3A). In contrast to ROCK2 overexpression, stable knockdown of ROCK2 (ShROCK2-1 and ShROCK2-3) considerably decreased cell migration ability in BEL7402 as evidenced by Transwell assay (P < 0.001, t test) (Fig. 3B) and wound healing assay (Fig. 3C). Of note, ROCK2 expression was almost completely knocked down in ShROCK2-1 and partially knocked down in ShROCK2-3 (Fig. 3A). Intriguingly, ShROCK2-1 displayed a drastic reduction of migrated cells compared with ShROCK2-3, and this suggested a possible dosage effect of ROCK2 on HCC cell migration (Fig. 3B). Consistently, ROCK2 stable knockdown transfectant (ShROCK2-1) also exhibited a significant decrease in the number of invaded cells compared with its vector control (P = 0.007, t test) (Fig. 3D). The results from the in vitro overexpression and knockdown cell models suggested that ROCK2 played a significant role in HCC cell migration and invasion.

Figure 3.

Knockdown of ROCK2 suppresses HCC cell migration and HCC cell invasion. (A) ROCK2 expression in ROCK2 stable knockdown cells. Stable clones were generated from the BEL7402 HCC cell line. Both ShROCK2-1 and ShROCK2-3 stable clones showed significant reduction of ROCK2 expression as compared with the vector stable clones V1 and V3. Protein bands were scanned and quantified using AlphaEaseFC software. (B) Transwell cell migration assay. 1 × 105 cells were seeded onto the upper chamber of the transwell and allowed to migrate for 16 hours. The number of migrated cells was significantly reduced in ROCK2 stable knockdown cells (ShROCK2-1 and ShROCK2-3) as compared with vector controls (V1 and V3) (P < 0.001, t test). (C) Wound healing assay. Wound closure was delayed in ROCK2 stable knockdown cells (ShROCK2-1) as compared with vector control (V3) in both 24- and 48-hour time points. (D) Invasion assay. 3 × 105 cells were seeded onto Matrigel-coated invasion chamber and allowed to invade for 24 hours. The number of invaded cells was significantly reduced in ROCK2 stable knockdown cells (ShROCK2-1) as compared with vector control (V3) (P = 0.007, t test). Error bars indicate the standard error of the mean.

Stable Knockdown of ROCK2 in HCC Cell Line Suppresses Invasion in Nude Mice.

To investigate the functions of ROCK2 on HCC cell invasion in vivo, orthotopic liver implantation was performed in nude mice with ROCK2 stable knockdown cell lines. To this end, the HCC cell line MHCC97L, which was shown to be invasive in an orthotopic liver implantation model,23 was sequentially transfected with short hairpin RNA targeting ROCK2 and firefly luciferase reporter. Stable knockdown of ROCK2 was confirmed via western blotting (Fig. 4A). After subcutaneous injection, tumors derived from MHCC97L ShROCK2 and vector clones were then resected and implanted into the livers of nude mice, with weekly monitoring of bioluminescent signals. After 6 weeks of implantation, the mice were sacrificed and their livers and lungs were examined. Pulmonary metastasis was found in one mouse from the vector group and none in the ShROCK2 group, as detected by the bioluminescence generated from the lung of the mouse and confirmed with histology (Fig. 4B,C; Table 2). A more remarkable effect in HCC local invasion was found from histological analysis of the tumors. In the orthotopic implanted livers, the tumor growth fronts from the MHCC97L vector group were invasive and irregular, whereas the tumor growth fronts from the MHCC97L ShROCK2 group were found to be bulging, more regular, and less invasive (Fig. 4D). In addition, tumor microsatellite formation was found in three of the four mice in the vector group but none in the ShROCK2 group (Fig. 4E, Table 2). Moreover, the incidence of venous invasion was slightly reduced in tumors derived from the MHCC97L ShROCK2 clone (Table 2). Of note, only a single focus of venous invasion was found in each of the two tumors of ShROCK2 clone, as compared with multiple foci of venous invasion found in the MHCC97L vector clone (Fig. 4F).

Figure 4.

Knockdown of ROCK2 in MHCC97L suppresses HCC invasion in vivo. (A) ROCK2 expression was significantly reduced in the ROCK2 stable knockdown HCC cell line, MHCC97L. (B) A luciferase signal was generated from luciferase-labeled MHCC97L xenografts. The signal was observed in the livers of the nude mice where the xenografts were implanted. A stronger signal was observed in the vector control group. One mouse from the vector control group displayed lung metastasis (white rectangular box). Lungs were dissected out and luciferase signal was measured separately. Positive luciferase signal was detected from the lung, which is indicative of lung metastasis. (C) Lung metastasis (arrow) found in the vector control group was confirmed via histological analysis with hematoxylin-eosin staining. (D) Tumors formed in livers from MHCC97L vector control clone displayed irregular invasive tumor fronts, whereas tumors formed in livers from the MHCC97L ShROCK2 clone displayed bulging growth fronts (hematoxylin-eosin sections). Arrows mark the tumor fronts. (E) Tumor microsatellite formation (arrows) around the main tumor was observed in the mice of the vector control group but not in those of the ShROCK2 group. (F) Multiple foci of venous invasion (arrows) were detected in the livers of the MHCC97L vector control group, whereas only a single focus of venous invasion (arrow) was detected in each of the two livers of the MHCC97L ShROCK2 group.

Table 2. Results of Orthotopic Liver Implantation from MHCC97L Vector and ROCK2 Knockdown Clones
 No. of Tumors with Invasive Growth FrontNo. of Tumors with Venous InvasionNo. of Tumors with Microsatellite FormationNo. of Tumors with Lung Metastasis
Vector4/43/4 (multiple foci in all)3/41/4
ShROCK21/42/4 (single focus in both)0/40/4

We also repeated similar orthotopic liver implantation experiments using subcutaneous tumor xenografts derived from a BEL7402 vector clone (V3) and BEL7402 ROCK2 knockdown ROCK2 clone (ShROCK2-1). Similarly, tumors formed from BEL7402 vector control clone displayed frequent venous invasion than that of BEL7402 ShROCK2 clone (ShROCK2-1) (Supplementary Fig. 3). Overall, we demonstrated that stable knockdown of ROCK2 was able to suppress HCC invasion both in vitro and in vivo.

In the livers of mice, we observed a slight reduction of tumor size derived from MHCC97L ShROCK2 cell line as compared with the MHCC97L vector control cell line. Nevertheless, no difference was observed in the in vitro cell proliferation rate (Supplementary Fig. 4A,B, left panels) and in vivo subcutaneous tumor size when ROCK2 was knocked down in both HCC cell lines (Supplementary Fig. 4A,B, right panels). Although the precise reason needs to be elucidated, these observations could be explained by the disparate growth environments, and we speculate that ROCK2 may also be important to the tumor development of HCC cells in the mice's livers that mimic the natural microenvironment of HCC growth in humans.

Knockdown of ROCK2 Inhibits Stress Fiber Formation and Attenuates Phosphorylation of Myosin Phosphatase Target Subunit.

To investigate the molecular mechanisms of ROCK2 in regulating cell motility and cancer invasion, we examined stress fiber formation and polymerized actin in ROCK2 stable knockdown cells. Cytoskeletal reorganization exemplified by the formation of stress fiber bundling arrays is essential for the contractile motion of cancer cells. Using phalloidin staining, we found that stress fiber formation was suppressed in ROCK2 stable knockdown transfectant ShROCK2-1 (Fig. 5A).

Figure 5.

Knockdown of ROCK2 impairs cytoskeletal events. BEL7402 ROCK2 knockdown stable clone and vector control stable clone were used. (A) Knockdown of ROCK2 disrupted stress fiber network. Stress fibers (polymerized actin) and actin filaments were demonstrated by phalloidin staining (green). Top panel: vector control transfectant (V3) of BEL7402 cells exhibited clear stress fibers, whereas ROCK2 knockdown stable transfectant (ShROCK2-1) exhibited loss of stress fibers (fluorescence microscopy; magnification ×1,000). (B) Knockdown of ROCK2 inhibited MYPT phosphorylation at Thr 853. With western blot analysis, ShROCK2 stable transfectant (ShROCK2-1) displayed a reduction in MYPT phosphorylation (Thr 853), indicative of impairment of actomyosin contractility. (C) Knockdown of ROCK2 suppressed formation of filopodia and lamellipodia (arrows) (scanning electron microscope; magnification ×5,000 [left panel] and magnification ×8,000 [right panel]).

Previous studies have indicated that phosphorylation of myosin phosphatase target subunit (MYPT1) at Thr 853 is essential for phosphorylation of myosin II and crucial for cell contractility.24 To further delineate the roles of ROCK2 on actomyosin contractility, the MYPT1 phosphorylation level in ROCK2 stable knockdown transfectant ShROCK2-1 was assessed. Using western blotting, we demonstrated that the phosphorylation level of MYPT1 (Thr 853) of ROCK2 knockdown stable transfectant was significantly reduced compared with the vector control (Fig. 5B), suggesting that ROCK2 is important for mediating cytoskeletal reorganization in HCC cells via regulating actomyosin contraction.

Knockdown of ROCK2 Inhibits Filopodia and Lamellipodia Formation.

To further understand the impact of ROCK2-mediated cytoskeletal reorganization on migratory and invasive related cellular changes in HCC cells, we compared the cell morphology of vector and ShROCK-2 stably transfected BEL7402 cells. Under a scanning electron microscope, we observed a significant reduction of filopodia and lamellipodia (cell protrusions) on the cell surfaces of ROCK knockdown cells compared with the vector control cells (Fig. 5C). Filopodia and lamellipodia are dynamic cellular features on cell membrane surfaces, require actin polymerization, and are involved in cancer cell invasion.25, 26 Our results indicate that ROCK2 might play an important role in regulating the cellular events related to cancer invasion.

Knockdown of ROCK2 Disrupted HCC Cell Directional Movement.

Next, we investigated the effect of ROCK2 on HCC cell directional movement via Golgi reorientation assay.22 Migrating cell coordinates directional movement through protruding at the anterior end of the membrane and retracting from the posterior end of the cell. Directional movement also plays an important role in cancer invasion and metastasis. Cell directional movement can be shown via realignment of the Golgi apparatus to the front of the nucleus in the direction of movement in order to direct membrane protrusion.22, 27 Therefore, we examined the effect of ROCK2 in directional movement by comparing the abilities of BEL7402 V3 and BEL7402 ShROCK2-1 cells in Golgi apparatus repositioning. At zero time point, the random values of cells scoring positive for Golgi reorientation assay in both groups were 15% (Fig. 6A). Five hours after the wound wascreated, 55% of the cells from the BEL7402 V3 clone scored positive, compared with 28% cells from the BEL7402 ShROCK2-1 cells (P = 0.001) (Fig. 6A,B). Our results show that knockdown of ROCK2 significantly impairs the ability of HCC cells for directional movement.

Figure 6.

Knockdown of ROCK2 suppresses HCC cell directional movement. (A) Percentage of cells exhibited the repositioning of the Golgi to the front of the nucleus toward the direction of movement at a given time point after wounds were created. A range of 353 to 737 cells in total were counted for each column. Error bars represent the standard error of the mean. (B) Representative pictures of Golgi repositioning 5 hours after the wound was created in BEL7402 vector control transfectant (V3) (top panel) and BEL7402 ShROCK2 stable transfectant (ShROCK2-1) (bottom panel). The Golgi apparatus was stained with lectin HPA Alexa Fluor 488 conjugates (green), and the nuclei were stained with 4',6-diamidino-2-phenylindole (blue). The wounds are marked with white lines, and arrows indicate the direction of cell movement. The BEL7402 ShROCK2 stable transfectant (ShROCK2-1) showed significant reduction in the realignment of the Golgi apparatus to the front of the nuclei at the direction toward the wound, as compared with the BEL7402 vector control transfectant (V3) 5 hours after the wound was created (P = 0.001, t test).


In this study, we found that ROCK2 protein was frequently overexpressed in human HCC. Intriguingly, with real-time quantitative PCR performed on the same 22 pairs of HCC cases with ROCK2 protein overexpression, ROCK2 messenger RNA was found to be unchanged between tumorous and nontumorous livers (Supplementary Fig. 5). This finding suggests that ROCK2 expression is deregulated in HCC posttranscriptionally, and further understanding about the mechanisms involved is needed. Importantly, overexpression of ROCK2 protein significantly correlated with tumor microsatellite formation in human HCCs, which is a pathological feature of intrahepatic metastasis in HCC. Furthermore, we found that overexpression of ROCK2 promoted HCC cell motility and invasiveness of HCC cells in vitro. This mirrored the findings of ROCK2 overexpression in human HCC samples and its association with more frequent tumor microsatellite formation. Conversely, when we knocked down the endogenous ROCK2 specifically using a short hairpin RNA approach in HCC cells, the mobility and invasiveness of HCC cells were significantly reduced, indicating that ROCK2 is closely involved in HCC invasion. Importantly, our in vivo animal model using orthotopic liver implantation also confirmed that ROCK2 was associated with tumor invasion. The specific roles of ROCK2 in HCC have never been addressed, so our findings are significant. Reduction of ROCK2 expression might be a potential target for antagonizing tumor invasion in HCC.

Studies have shown that inhibition of Rho-kinase activity by ROCK specific inhibitor Y27632 could suppress HCC cell migration, invasiveness, and intrahepatic metastasis of HCC,15–17 as well as tumor recurrence after liver transplantation.28 Nevertheless, Y27632 inhibits the kinase activity of both ROCK1 and ROCK2 and even other kinases; therefore, ROCK inhibitor does not specifically distinguish the contributions of ROCK1 and ROCK2 in hepatocarcinogenesis. From the parallel examination of ROCK1 and ROCK2 protein expression in human HCC, we found the particular significance of ROCK2 in human HCC. This would shed light on the development of more specific anti-HCC drugs that target ROCK2 exclusively.

ROCK1 and ROCK2 share a 65% sequence homology in amino acid level.7 Although both ROCK1 and ROCK2 were identified as the downstream targets of RhoA,29, 30 studies have confirmed that they are regulated and processed by different participants and are not functionally redundant.31, 33 Knockout mice models have provided significant evidence distinguishing their individual physiological roles. ROCK1 and ROCK2 homozygous knockout (ROCK1−/− and ROCK2−/−, respectively) mice displayed distinct phenotypes. ROCK1−/− mice manifested EOB (eyelid open at birth) and omphalocele phenotypes,4 whilst ROCK2−/− mice manifested not only EOB, but also thrombus formation, placental dysfunction, and intrauterine growth retardation.5 Furthermore, most ROCK2−/− mice were embryonic fatal,5 whereas most ROCK1−/− mice could survive till adulthood.4 On top of this growing body of evidence, the present study added a new layer of evidence showing a distinctive role of ROCK2 in human HCC. This indicated that, of the two members of the Rho-kinase family, ROCK2 might play a more important role in hepatocarcinogenesis.

Many studies have evidenced and depicted the roles of Rho/ROCK pathway in the regulation of cytoskeletal events. The Rho/ROCK pathway is indispensable for cell movement, including the coordination of actomyosin contraction,33 rear end retraction,34 turnover of focal adhesions, and formation of stress fibers in fibroblasts.35 Our data substantiated that ROCK2 was a key mediator regulating the phosphorylation of MYPT1 and thereby actomyosin contraction. We further demonstrated that specific knockdown of ROCK2 in BEL7402 significantly suppressed the formations of stress fiber, filopodia, lamellipodia, and cell directional movement. A cell moves to a designated direction with cooperation of continuous actin polymerization and depolymerization that allow the cell to protrude at the anterior front, undergo consecutive actomyosin contractions, and de-adhere from the posterior end. Formations of stress fiber, filopodia, and lamellipodia require actin polymerization, and loss of these features is indicative of the cell's inability in actin polymerization. Altogether, our data demonstrate that ROCK2 is important for the multiple events in HCC cell migration.

Acquisition of cell migratory and invasive abilities is essential for cancer invasion and metastasis, which involve complicated changes in gene expression. In addition to ROCK2 expression, different genes have been reported to be deregulated in HCC, and such deregulations contribute to HCC invasion. A group of proteins called matrix metalloproteinases (MMPs) were shown to be involved in cancer invasion and metastasis. Overexpression of MMP2 and MMP9 in HCC was shown to be involved in increased tumor recurrence or metastasis after tumor resection.36 Particularly, MMP9 was demonstrated to increase the invasive potential of HCC and was involved in capsular filtration in HCC.37 Osteopontin, a secreted phosphoglycoprotein that binds to integrin receptors, was shown to be frequently overexpressed in HCC. It has recently been shown to be involved in in vitro invasion and in vivo lung metastasis through the up-regulation of MMP2 and a serine proteinase, urokinase.38 Furthermore, TWIST, a transcriptional factor, was found to correlate with HCC metastasis through down-regulation of E-cadherin.39 These reported proteins are known to be important for various steps involved in HCC metastasis, such as invasion of the extracellular matrix and loss of cell–cell adhesion. In this study, we report that ROCK2 was overexpressed in human HCCs and this overexpression was associated with a more aggressive biological behavior. Our results show that ROCK2 played a significant role in HCC progression and invasion by regulating cytoskeletal reorganizations and enhancing the invasiveness and motility of HCC cells both in vitro and in vivo.


We thank Q. Cheng, D. Y. Guo, and Kevin Ng for help in the orthotopic liver implantation model; D. Y. Jin and K. H. Kok for valuable advice and discussion; and Joyce M. F. Lee for help in the immunohistochemistry study.