Suppression of actopaxin impairs hepatocellular carcinoma metastasis through modulation of cell migration and invasion

Authors

  • Lui Ng,

    1. Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Ronnie Tung-Ping Poon,

    1. Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    2. Centre for Cancer Research, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Simon Yau,

    1. Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Ariel Chow,

    1. Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    2. Centre for Cancer Research, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Colin Lam,

    1. Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Hung-Sing Li,

    1. Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Thomas Chung-Cheung Yau,

    1. Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    2. Centre for Cancer Research, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Wai-Lun Law,

    1. Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    Search for more papers by this author
  • Roberta Pang

    Corresponding author
    1. Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    2. Centre for Cancer Research, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong
    • Address reprint requests to: Dr Roberta Pang, Ph.D., Department of Surgery, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong. E-mail: robertap@hku.hk; fax: +852-28173903.

    Search for more papers by this author

  • Supported by the Central Allocation Group Research Grant “Molecular Pathology of Liver Cancer — A Multidisciplinary Study” from the Research Grant Council of Hong Kong (HKU-7-CRF09).

  • Potential conflict of interest: Nothing to report.

Abstract

Early reports suggested that actopaxin, a member of the focal adhesion proteins, regulates cell migration. Here we investigated whether actopaxin is involved in hepatocellular carcinoma (HCC) progression and metastasis. We examined actopaxin expression in human HCC samples using immunohistochemistry and western blotting. The functional and molecular effect of actopaxin was studied in vitro by overexpression in a nonmetastatic HCC cell line, as well as repression in a metastatic cell line. The in vivo effect of actopaxin repression was studied in nonobese diabetic and severe combined immunodeficient mice. We found that actopaxin was frequently overexpressed in human HCC patients and its overexpression positively correlated with tumor size, stage, and metastasis. Actopaxin expression also correlated with the metastatic potential of HCC cell lines. Actopaxin overexpression induced the invasion and migration ability of nonmetastatic HCC cells, whereas down-regulation of actopaxin reverted the invasive phenotypes and metastatic potential of metastatic HCC cells through regulating the protein expression of certain focal adhesion proteins including ILK, PINCH, paxillin, and cdc42, as well as regulating the epithelial-mesenchymal transition pathway. Furthermore, there was a close association between actopaxin and CD29. HCC cells with stronger CD29 expression showed a higher actopaxin level, whereas actopaxin repression attenuated CD29 activity. Finally, actopaxin down-regulation enhanced the chemosensitivity of HCC cells towards oxaliplatin treatment by way of a collective result of suppression of survivin protein, β-catenin, and mammalian target of rapamycin pathways and up-regulation of p53. Conclusion: This study provides concrete evidence of a significant role of actopaxin in HCC progression and metastasis, by way of regulation of cell invasiveness and motility, an epithelial-mesenchymal transition process, and chemosensitivity to cytotoxic drugs. (Hepatology 2013;58:667-679)

Abbreviations
97L-LUC cells

luciferase incorporated MHCC97L cells

ECM

extracellular matrix

H&E

hematoxylin and eosin

HCC

hepatocellular carcinoma

IHC

immunohistochemistry

NOD-SCID

nonobese diabetic and severe combined immunodeficient

Hepatocellular carcinoma (HCC) is the third leading cause of cancer death worldwide.[1] Most HCC patients die from locally advanced or metastatic disease in a relatively short period of time, and the mechanisms responsible for HCC progression and metastasis remain a major challenge to researchers in this field. It is believed that the elucidation of molecular mechanisms involved in HCC progression and metastasis is important for development of HCC therapeutic agents.

Tumor invasion and metastasis are complex processes requiring the ability of tumor cells to interact with the extracellular matrix. Major cell surface receptors that mediate these interactions are integrins.[2] Several studies have demonstrated the important roles of β1-integrin in the growth, invasiveness, and metastatic potential of HCC,[3-5] as well as in protecting cancer cells against chemotherapeutic-drug-induced apoptosis.[6] Focal adhesion proteins serve as a structural and signaling nexus connecting integrins and the dynamic actin cytoskeleton to coordinate cell motility and cell invasion.[7] As such, dysregulation of focal adhesion proteins very often contribute to the acquired metastatic behavior of cancer cells.

Actopaxin, a focal adhesion protein, associates with many other focal adhesion molecules to regulate cellular events including cell adhesion and spreading[8, 9] as well as cell migration.[10-12] More important, its phosphorylation has been recently shown to induce matrix degradation, cell migration, and invasion in breast cancer cell line models.[13] Yet the clinical significance of actopaxin has not been demonstrated so far.

In this study we determined the expression of actopaxin in clinical HCC specimens and studied the in vitro and in vivo functions of actopaxin on HCC invasion and metastasis.

Materials and Methods

Patients and Specimens

Fresh tumor specimens were obtained with informed consent from patients who underwent surgical resection of primary HCC at the Department of Surgery, Queen Mary Hospital, the University of Hong Kong. The study was approved by the Institutional Review Board.

Cell Lines, Tissue Culture, Transfections, and Reagents

Transfections of actopaxin small interfering RNA (siRNA) (purchased from Invitrogen, La Jolla, CA) and CD29 siRNA (Purchased from Qiagen, Valencia, CA) at a final concentration of 50 nM was performed using Lipofectamine 2000 reagent (Invitrogen). The universal negative control (Invitrogen) and AllStarts siRNA (Qiagen) was used as control. For western blotting, cells were analyzed at day 3. For downstream experiments using siRNA transfected cells, cells were transfected for 6 hours then recovered in complete medium without antibiotics for 2 hours.

Plasmids for stable knockdown of actopaxin were purchased commercially (Origene, Rockville, MD) where the short hairpin RNA (shRNA) was cloned into a pGFP-V-RS vector under U6 promoter according to our provided actopaxin siRNA sequence (CCAGGAGCAUCAAGUGGAAUGUGGA). Stable transfections of actopaxin-shRNA and the negative control plasmid (Origene) into 97L-LUC cells were performed using Lipofectamine 2000 reagent (Invitrogen).

Invasion and Migration Assays

Invasion and migration assays were performed in 24-well Matrigel-coated filers (8.0 μm pore size, BD) and PET inserts (8.0 μm pore size, Corning), respectively, according to the manufacturer's instructions. Briefly, siRNA-transfected or shRNA cells (1 × 105) were plated in Dulbecco's modified Eagle's medium (DMEM) into invasion and migration chambers and the lower part of the chamber was filled with DMEM containing 10% fetal bovine serum (FBS) as chemoattractant. After 16 hours, cells on the upper part of the membrane were removed with a cotton swab and invaded/migrated cells were fixed in cold methanol, stained with 0.1% crystal violet, and counted.

Isolation of CD29 Strong and Weak Subpopulations

Harvested 97L-LUC cells or dissociated clinical HCC (THCC) and adjacent nontumorous liver cells were incubated with a monoclonal PE conjucated-CD29 antibody (BioLegend, Montgomery, TX) for 30 minutes at 4°C and sorted on a MoFlo XDP Cell Sorter (Beckman Coulter) with immunofluorescence data analyzed on the Summit Software v. 5.0 (Beckman Coulter). The purity of the isolated subpopulations regularly exceeded 90%. All fluorescent-activated cell sorter (FACS) analyses and sorting were paired with the matched isotype control (BioLegend, San Diego, CA).

Protein Extraction and Western Blot Analysis

Cell lines were lysed in RIPA buffer (Cell Signaling Technology, Danvers, MA) containing 1 mmol/L phenoylmethylsulfonyl fluoride for 1 hour on ice. For western immunoblotting of clinical specimens, HCC tissues were homogenized in lysis buffer with 2.0 mol/L urea. Lysed protein was suspended in sodium dodecyl sulfate sample buffer, boiled, resolved in sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and transferred to PVDF membranes (GE Healthcare, Piscataway, NJ). Antibodies against actopaxin, cdc42, E-cadherin, GSK3β, cyclin D1, phospho mammalian target of rapamycin (mTOR) (ser2448), phospho 4E-BP1 (Thr37/46), and phospho p53 (ser15) were purchased from Cell Signaling Technology. Anti-actin was from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies against ILK, PINCH, and β-catenin were from BD Biosciences (San Diego, CA). Anti-paxillin was from Calbiochem (La Jolla, CA). Anti-survivin was from Norvus Biologicals (Littleton, CO). Representative western blots of experiments that were repeated 3-4 times are shown in every case.

Animal Work

The protocol was approved by the Committee on the Use of Live Animals in Teaching and Research (CULATR) of the University of Hong Kong. Tumors were allowed to grow in nonobese diabetic and severe combined immunodeficient (NOD-SCID) mice by injection of 97L-LUC shCTL and shRNA stable cells either subcutaneously into flank regions or orthotopically onto liver with 1 × 106 cells per site. At week 6 postoperation, luciferase signal was generated by intraperitoneal injection of luciferin substrate and captured with the Xenogen IVIS100 System.

Immunohistochemical Study of Clinical and Mice Specimens

Formalin-fixed and paraffin-embedded specimens were cut into 5-μm-thick sections by microtome and mounted on slides coated with 3-aminopropryltriethoxysilane (TESPA) in acetone. Sections were then deparaffinized in three changes of xylene and rehydrated in serial dilutions of ethanol. Antigen retrieval was performed by microwave treatment at low power for 10 minutes in preheated citrate buffer. Endogenous peroxidase and endogenous biotin activity were blocked using a biotin blocking kit, then blocked with horse serum for 30 minutes. Sections were incubated with the primary antibodies anti-actopaxin (Abgent, San Diego, CA) and anti-Ki67 (Dako, Carpinteria, CA) at 1:100 dilutions overnight at 4°C in a moist chamber. After incubation, slides were rinsed three times with TBS-Tween and twice with TBS, and probed with biotinylated secondary antibody for an hour. The slides were then rinsed three times with TBS-Tween and twice with TBS and probed with avidin-horseradish peroxidase (HRP) for an hour. Sites of bound antibody were visualized using liquid DAB+substrate-chromogen system. Finally, sections were counterstained with Gill's hematoxylin and mounted in DPX mountant (BDH Laboratory, UK).

Cell Viability Assay

The same number of siRNA-transfected and shRNA stable cells were seeded on 96-well plates for 24 hours and then subjected to commonly used chemotherapeutic oxaliplatin at low (2.5 μM) and median concentrations (5 μM). The cell viability at 72 hours after drug treatment were then determined using 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide (MTT) (Invitrogen).

Cell Cycle Analysis

The same number of siRNA-transfected and shRNA stable cells lines were seeded on 6-well plate for 24 hours and then treated with oxaliplatin. After 48 hours cells were collected and fixed in ice-cold 70% ethanol at 4°C overnight. Cells were then pelleted by centrifugation and rehydrated in phosphate-buffered saline (PBS) at room temperature for 10 minutes, followed by staining with propidium iodide/RNase A (Invitrogen) at 37°C for 30 minutes. The stained cells were analyzed by CytomicsFC500 Flow Cytometry (Beckman Coulter) and the results were analyzed using the ModFit LT2.0 software.

Apoptotic Assay

Cells were seeded and treated as described in the Cell Cycle Analysis section. After 48 hours cells were collected and subjected to apoptotic assay using the PE Annexin V Apoptosis Detection Kit (BD Pharmingen) according to the manufacturer's protocol.

Statistical Analysis

Data analysis was performed using SigmaStat 3.5 (Systat Software, San Jose, CA) or SPSS 16.0 (Chicago, IL). The Mann-Whitney test and Student t test were used to analyze differences between experimental groups of clinical specimens and cell line models, respectively. Spearman's correlation test was applied to determine correlations. Disease-free survival (DFS) and overall survival (OS) were calculated by the Kaplan-Meier method, and differences in survival rate were compared using the log-rank test. Categorized data were analyzed with Fisher's exact test. Univariate and multivariate analyses were performed using the Cox proportional hazards regression model. P < 0.05 was considered statistically significant.

Results

Elevated Actopaxin Expression and Its Association With Tumor Size, Metastatic Potential, and Survival of HCC Patients

Immunohistochemistry (IHC) was first performed to examine the expression pattern of actopaxin in 32 pairs of HCC specimens and adjacent nontumorous liver. Fifteen of them (∼47%) displayed stronger cytoplasmic staining in the tumor specimens when compared with the paired nontumorous livers (Fig. 1A). To precisely quantify the fold of actopaxin overexpression, western blot analysis was further applied to detect the actopaxin expression in another set of clinical HCC specimens and the paired adjacent nontumorous livers from 75 patients, in which 68 underwent curative resection and seven underwent liver transplantation. The relative expression of actopaxin was higher in HCC specimens when compared with their adjacent nontumorous livers (P = 0.023) (Fig. 1B). These results collectively showed that actopaxin overexpression was frequently observed in HCC patients.

Figure 1.

Actopaxin expression correlated with HCC progression and metastasis. (A) A representative IHC staining showed that actopaxin was frequently overexpressed in HCC specimens (THCC) when compared with the adjacent nontumorous liver (NHCC) (original magnification 200×). (B) Relative expression of actopaxin in 75 HCC patients determined by western blot analysis was significantly higher in THCC than that in the paired NHCC (right panel). One representative western blotting is shown (left panel). (C) Fold of actopaxin overexpression positively correlated with increasing tumor size. (D) Fold of actopaxin overexpression was significantly higher in HCC samples of higher stage (left panel) and with metastatic HCC (right panel). (E) Correlation of actopaxin overexpression with disease-free and overall survival rates of patients. (F) Actopaxin expression in nonmetastatic and metastatic HCC cell lines.

The clinicopathological significance of actopaxin overexpression in HCC was determined (Table 1). Actopaxin overexpression positively correlated with tumor size (P = 0.029; Fig. 1C) and stage (P = 0.013; Fig. 1D). Moreover, overexpression of actopaxin associated with development of extrahepatic metastasis in patients within 10 years after surgery. Over 50% of patients (15 of 29) with metastatic HCC showed at least a two-fold actopaxin overexpression, whereas such overexpression was detected in only ∼26% (12 of 45) of patients with nonmetastatic HCC (P = 0.047). The degree of actopaxin overexpression in metastatic HCC patients was also significantly higher than that in the nonmetastatic patient group (P = 0.012; Fig. 1D). Furthermore, the prognostic value of actopaxin overexpression in patients who underwent hepatic resection was determined. Patients with lower actopaxin expression showed a significantly longer DFS (141.7 versus 58.4 months, P < 0.001) and OS (123.1 versus 70.1 months, P = 0.002) when compared with those with higher actopaxin expression (Fig. 1E). Our univariate analysis revealed that actopaxin expression, tumor size, AJCC stage, and microvascular invasion were significantly correlated with DFS and OS (Supporting Table 1). These four clinicopathological parameters were further applied for multivariate analysis, and the results showed that actopaxin overexpression, tumor size, and microvascular invasion were independent risk factors for DFS, while actopaxin overexpression, AJCC stage, and microvascular invasion were independent risk factors for OS (Supporting Table 1). Our statistical analysis revealed the prognostic significance of actopaxin overexpression for DFS and OS of HCC patients.

Table 1. Summary of Findings of Clinicopathological Correlation
Clinicopathological FeaturesCategoryNumber of CasesaActopaxin Overexpression THCC/NHCC (25%/Median/75%)P Value
  1. a

    In some categories the total number of cases is less than 75 due to incomplete information.

  2. b

    No correlation between actopaxin expression and tumor size was observed using the conventional tumor size cutoff (5 cm), but showed significant correlation when the cutoff increased to 7 cm.

Age<50310.75/1.16/2.540.273
 > or = 50430.92/1.29/5.86 
GenderMale590.92/1.50/5.150.079
 Female150.58/0.99/1.59 
HBsAgNegative100.58/0.99/11.990.472
 Positive610.90/1.28/3.65 
Microvascular invasionAbsent480.85/1.16/3.180.623
 Present270.94/1.50/4.55 
Satellite noduleAbsent520.92/1.26/5.460.635
 Present190.83/1.19/3.26 
CirrhosisAbsent460.59/1.15/2.920.205
 Present270.96/1.26/6.28 
Tumor sizeb<7 cm350.70/1.07/2.000.018
 >7 cm391.08/1.53/10.42 
Tumor AJCC stageI to II370.72/1.10/2.140.013
 III to IV321.10/1.61/17.65 
Distant metastasisAbsent460.72/1.15/1.990.018
 Present291.04/1.95/17.23 

Actopaxin Overexpression Induces Cell Invasion and Migration of HCC

The association of actopaxin expression with HCC metastatic potential was supported by protein expression of actopaxin in a series of nonmetastatic and metastatic HCC cell lines (Fig. 1F). Compared with the nonmetastatic cells Hep3B, H2P, Huh7, and PLC/PRF/5, the actopaxin expression was higher in the metastatic cells H2M, MHCC97L, and MHCC97H. In addition, H2M, which is the metastatic form of H2P, and 97H, which is more metastatic than 97L, showed higher actopaxin protein expression.

We next examined whether actopaxin overexpression in nonmetastatic HCC cells induced their metastatic behavior via stable transfection of pcDNA-actopaxin construct into PLC cells (Fig. 2A). The ability of the stable cells to invade and migrate was determined by in vitro assays. The number of invaded cells was significantly higher in stable PLC-actopaxin transfectants (Fig. 2B), with ∼540% increase in the number of invaded cells when compared with the vector control. Overexpression of actopaxin also resulted in a marked increase in cell motility (Fig. 2C); the number of stable PLC-actopaxin transfectants migrated was 280% higher than the vector control.

Figure 2.

Actopaxin overexpression induced HCC cell invasion and migration. (A) Expression of actopaxin in nonmetastatic HCC cells PLC stably transfected with pcDNA (vector) or pcDNA-actopaxin (actopaxin). (B) Actopaxin stable transfectant showed enhanced invasion potential when compared with the vector control. (C) Actopaxin stable transfectant showed enhanced migration potential.

Actopaxin Repression Attenuates HCC Cell Invasion and Migration

To further validate the functional significance of actopaxin expression in HCC, actopaxin expression in metastatic cell line 97L-LUC was suppressed by RNA interference. For stable suppression, actopaxin expression was inhibited by shRNA transfected into 97L-LUC cells, whereas transient down-regulation was achieved by actopaxin siRNA duplexes. The efficacy of actopaxin repression was confirmed by western blotting (Fig. 3A).

Figure 3.

Actopaxin repression impaired HCC cell invasion and migration. (A) Actopaxin expression in 97L-LUC stable actopaxin shRNA clones (left panel). Pooled colonies stably transfected with the vector served as the negative control (shCTL). Expression of actopaxin in siRNA or negative control (siCTL) transient transfected 97L-LUC cells (right panel). (B) A reduced number of invaded cells was observed in actopaxin shRNA clones (left panel) and siRNA repressed cells (right panel). (C) A reduced number of migrated cells was observed in actopaxin shRNA clones (left panel) and siRNA repressed cells (right panel).

The in vitro effects of actopaxin knockdown on HCC cell invasion and migration of 97L-LUC was examined. Actopaxin repression significantly impaired the invasiveness of 97L-LUC cells (Fig. 3B). Actopaxin shRNA transfectants shRNA, shRNA4, and shRNA5 showed fewer invaded cells (5%, 48%, and 78%, respectively) when compared with the shCTL. In addition, among the shRNA transfectants the number of invaded cells was associated with their actopaxin expression level. Similarly, the number of siRNA cells invaded was ∼40% of siCTL transfectants (Fig. 3B). These results suggested that down-regulation of actopaxin expression accordingly suppressed the invasiveness of metastatic HCC cells. Down-regulation of actopaxin also resulted in a marked decrease in cell motility (Fig. 3C). Actopaxin shRNA transfectants shRNA, shRNA4, and shRNA5 showed fewer migrated cells (15%, 25%, and 45%, respectively) when compared with the shCTL, and the number of migrated cells was associated with their actopaxin expression level. Similarly, the number of migratory transient actopaxin siRNA transfectants was also markedly decreased, which was ∼55% of the siCTL transfectants.

Actopaxin Protein Expression Associates With Protein Expression Patterns of Higher Metastatic Potential

The regulatory effect of actopaxin expression on cell invasion and migration was associated with altered levels of certain focal adhesion proteins and the epithelial-mesenchymal transition (EMT) pathway. Actopaxin stable transfectants showed increased levels of focal adhesion proteins including ILK and PINCH (Fig. 4A), whereas stable and transient actopaxin down-regulation repressed their expressions (Fig. 4B,C). Actopaxin form ternary complex with ILK and PINCH, which is conserved in organisms ranging from nematodes and insects to humans,[14] and the formation of this complex is important for the stability of these proteins in cells. We found in the co-immunoprecipitation (co-IP) experiment that actopaxin forms complex with ILK and PINCH in HCC cells (Fig. 4D). Moreover, the moderate expressions of ILK and PINCH following actopaxin down-regulation suggested that actopaxin plays a predominant role in regulating HCC cell invasion and migration independent of ILK and PINCH expression. In addition, the expression of paxillin and cdc42, which correlated with lamellipodial protrusion and directed cell movement,[15, 16] was also positively associated with altered actopaxin expression (Fig. 4A-C).

Figure 4.

Molecular mechanism associated with altered actopaxin expression. (A) Expressions of certain focal adhesion proteins and EMT regulators in PLC actopaxin stable transfectants and vector control. (B) Expressions of certain focal adhesion proteins and EMT regulators in 97L-LUC actopaxin shRNA stable clone. (C) Down-regulation of ILK, PINCH, paxillin, and cdc42 following actopaxin siRNA transfection in 97L-LUC cells. (D) Co-IP experiment demonstrating presence of actopaxin in 97L-LUC cell lysate immunoprecipitated by anti-ILK and -PINCH antibodies (upper and lower panels, respectively). (E) Subcellular fractionation experiment showing overexpression of β-catenin was only observed in membranous and cytoskeletal fractions (upper panel), and co-IP experiment demonstrating enhanced E-cadherin-β-catenin interaction in actopaxin down-regulated cells (lower panel). (F) Actopaxin regulatory pathways on cell migration, EMT, invasion, and cancer metastasis.

More important, altered actopaxin expression in stable clones regulated the expressions of proteins involved in the EMT process. Compared with the vector control transfectant, actopaxin stable clone expressed lower level of epithelial cell markers E-cadherin and β-catenin (Fig. 4A). Such an effect was reversed in the actopaxin shRNA transfectant (Fig. 4B), suggesting that the metastatic cell 97L-LUC underwent the mesenchymal-epithelial transition following actopaxin repression. The regulatory effect of actopaxin on E-cadherin was associated with altered expression of Snail, GSK3β, and phospho-GSK3β(ser9) (Fig. 4B). Reduced expression of Snail, the transcription suppressor of E-cadherin, was detected in actopaxin shRNA cells. On the other hand, the expression of GSK3β, which regulates the nuclear export and degradation of Snail,[17] was strongly induced, whereas the level of phospho-GSK3β(ser9), which is kinase-inactive, was decreased. Stable overexpression of actopaxin also resulted in increased expression of phospho-GSK3β(ser9) (Fig. 4A), yet we did not find the regulatory effect of GSK3β and Snail (data not shown).

Besides, in the shRNA clone elevated β-catenin expression was restricted to their membranous and cytoskeletal fractions but not in cytoplasmic and nuclear fractions, and co-IP experiments demonstrated higher levels of β-catenin/E-cadherin complexes in shRNA cells when compared with shCTL cells (Fig. 4E), indicating that stable actopaxin repression enhanced cell-cell adhesion, which contributed to a less invasive phenotype.

Actopaxin Correlated With CD29 Activity

CD29 (β1-integrin) plays important roles in the growth, invasiveness, and metastatic potential of HCC.[3-5] In response to the external environment, CD29 recruits a network of adaptor and signaling proteins to its cytoplasmic domain and regulates their expressions and thus many cellular activities. We hypothesize that the actopaxin level is regulated by CD29 in this manner. To test our hypothesis, we first examined whether CD29 and actopaxin expressions were statistically correlated in HCC specimens. The CD29 protein expression in HCC specimens was determined by western blot analysis (Supporting Fig. 2) and showed significant positive correlation with the expression of actopaxin (P = 0.017, Fig. 5F). We next determined the regulatory effect of CD29 on actopaxin protein expression. Transient CD29 down-regulation by siRNA resulted in a decreased expression of actopaxin in 97L-LUC cells (Fig. 5A), yet the effect was not obvious due to the high endogenous CD29 expression. We thus applied the flow cytometric approach to sort 97L-LUC cells into CD29-strong and CD29-weak populations. Actopaxin expression was much higher in CD29-strong samples (Fig. 5B). These results were further supported by actopaxin expression in CD29-strong and CD29-weak dissociated HCC specimens (Fig. 5C), demonstrating that CD29 was an upstream regulator of actopaxin.

Figure 5.

Actopaxin expression associated with CD29 activity. (A) Actopaxin expression was reduced in 97L-LUC cells following transfection of CD29 siRNA duplexes for 72 hours. (B) 97L-LUC cells were sorted into CD29-strong and CD29-weak subpopulations using a flow cytometric approach. Actopaxin expression was higher in the CD29-strong subpopulation. (C) A representative result showed that actopaxin expression was higher in the CD29-strong subpopulation of dissociated clinical HCC specimens. (D) CD29 activity, which is demonstrated by phospho-CD29(thr788/789) level, was reduced in 97L-LUC actopaxin shRNA clone. (E) One representative IHC staining demonstrated that HCC specimens with actopaxin overexpression showed stronger phospho-CD29(thr788/789) staining in THCC. (F) Overexpression of actopaxin was significantly associated with CD29 and phospho-CD29(thr788/789) overexpression.

On the other hand, stable down-regulation of actopaxin did not affect the CD29 expression in 97L-LUC cells (Fig. 5D). Instead, we found that the phospho-CD29(thr788/789) level, which correlated with vascular invasion in HCC patients through alteration of CD29 extracellular conformation and the response to the external environment,[4] was decreased in actopaxin shRNA transfectant (Fig. 5D). Our IHC results further demonstrated that the phospho-CD29(thr788/789) level correlated with actopaxin overexpression in clinical specimens (P < 0.001; Fig. 5E,F). These results collectively demonstrated the association of actopaxin level with CD29 activity.

Actopaxin Repression Attenuates HCC Cell Metastasis In Vivo

The in vivo effect of actopaxin down-regulation on tumor cell metastasis was demonstrated in two animal models. In the first model, 97L-LUC shRNA and shCTL cells were subcutaneously injected into the flank regions of six NOD-SCID mice. Six weeks after injection, all mice developed tumors in the injected areas. Liver metastasis was observed in five out of six mice injected with shCTL cells, while none of the shRNA mice showed observable tumor in the liver (Fig. 6A). Moreover, in the shCTL-injected group, luciferase signal which was generated from 97L-LUC cells was detected in livers, lungs, and brains of all mice (Fig. 6B), indicating that the shCTL cells had metastasized from the flank regions to these sites. On the other hand, except that the brain of one mouse showed a weak luciferase signal, the luciferase signal was not observed in any organs of the shRNA-injected mice (Fig. 6B), indicating that actopaxin suppression impaired the metastatic potential of 97L-LUC cells.

Figure 6.

Actopaxin knockdown impaired metastasis in vivo. (A) Actopaxin shRNA and shCTL cells were injected subcutaneously into the flanks of NOD/SCID mice (1 million cells/site) and liver metastases were observed in the livers of all shCTL mice except mouse 3, whereas none of the shRNA mice showed development of liver metastasis. (B) Luciferase signals generated from injected 97L-LUC cells were detected from livers, lungs, and brains of all shCTL mice, whereas only one shRNA mouse showed weak luciferase signal in brain. (C) Actopaxin shRNA and shCTL cells were injected orthotopically into livers of NOD/SCID mice (1 million cells) and tumors were allowed to grow in liver for 6 weeks. All shCTL and shRNA mice showed development of tumor in livers. (D) Moderate to strong luciferase signal was generated from lungs of all shCTL mice, whereas weak luciferase signal was observed in two of the shRNA mice. (E) A representative H&E staining confirmed the presence of tumor development in livers and lungs of shCTL mice (left and right panels, respectively). Insets illustrate high-power magnification (400×) of the same field. (F) Development of tumors originated from 97L-LUC cells was further confirmed by IHC staining for human Ki67 antibody. Insets illustrate high-power magnification of the same field (400×). A representative section is shown.

In another model, actopaxin shCTL and shRNA stable cells were injected orthotopically into the liver of five NOD-SCID mice and the development of lung metastasis was examined. As shown in Fig. 6C, tumors grew on the livers of all mice. There was no statistical difference between the tumor sizes and Ki67 staining in shCTL and shRNA groups (P = 0.233 and 0.87, respectively; Supporting Fig. 3A-C), indicating that actopaxin repression did not affect tumor growth. This observation is in accordance with the results obtained from our in vitro cell proliferation experiments (Supporting Fig. 3D). Moderate to strong luciferase signals were detected in lungs of all NOD-SCID mice injected with shCTL cells (Fig. 6D). In the shRNA group, only two mice showed very weak luciferase signals in their lungs. The presence of tumor metastasis in livers and lungs of shCTL mice were further confirmed by hematoxylin and eosin (H&E) staining as well as immunostaining with antihuman Ki67 antibody (Fig. 6E,F), respectively. On the other hand, no metastasized tumors were observed in the shRNA group, as demonstrated by H&E staining and negative Ki67 staining. In addition, our results showed that the mean brain luciferase signal of shCTL mice (690494.4 units) was significantly higher (t test, P = 0.002) than that of shRNA mice (12651.4 units), indicating that a significantly higher number of shCTL cells were metastasized to the brain when compared to the shRNA cells, yet no brain metastasis could be confirmed histologically in either groups, possibly due to the metastasized cancer cells requiring a longer time to develop into solid tumors in the brain. In summary, our results showed that down-regulation of actopaxin reduced the metastatic ability of HCC in vivo.

Actopaxin Repression Enhances Chemosensitivity of Metastatic HCC Cell

It is generally believed that in HCCs there is a positive correlation between the chemoresistant rate and tumor progression; for example, vascular invasion, intrahepatic metastasis, and poor differentiation.[18] We examined the role played by actopaxin in chemosensitivity. Actopaxin shRNA and shCTL stable clones were exposed to a series of chemotherapy drugs for 72 hours. We found that actopaxin repression significantly increased the cytotoxicity of oxaliplatin (Fig. 7A). Our results were further supported by the inclusion of two additional actopaxin-shRNA stable clones (shRNA4 and shRNA5), which showed that the oxaliplatin-induced growth-inhibition effect was stronger in all the stable clones compared with the shCTL clone, and shRNA5, which showed the highest actopaxin expression among the shRNA clones, displayed the highest residual cell number (Fig. 7B). We also found that oxaliplatin exerts growth inhibitory effects on 97L-LUC cells through impairing cell cycle progression. Under 5-μM oxaliplatin treatment, the amount of shRNA stable cells arrested at G2 phase was increased by >300% (Fig. 7C), which was significantly higher than that of shCTL cells (44% increase). Similarly, the proportion of shRNA cells entering G0/G1 phase was decreased by 65%, which was also significantly higher than the 40% reduction observed in shCTL cells. The apoptotic effect of oxaliplatin on actopaxin shRNA and shCTL cells was also investigated. As shown in Fig. 7D, shRNA cells showed a higher proportion of apoptotic cells after 48-hour and 72-hour oxaliplatin treatments when compared with the shCTL cells, although only the effect under 2.5-μM oxaliplatin treatment was statistically significant. Our results were supported by the above experiments using actopaxin transiently repressed cells as models (Fig. 7B-D), strengthening the likelihood of targeting actopaxin in improving the chemosensitivity of HCC patients.

Figure 7.

Knockdown of actopaxin expression enhanced chemosensitization in HCC cells. (A) 97L-LUC actopaxin shRNA and shCTL cells were treated with low (Lo) and median (Med) concentrations of a series of cytotoxic drugs for 72 hours. The number of cells was determined by MTT assay and expressed as percentage of that under control treatment. When compared with shCTL cells, shRNA cells showed a significantly stronger response to low and median concentrations of oxaliplatin, and low and median concentrations of 5-FU. (B) All actopaxin shRNA stable clones including shRNA, shRNA4, and shRNA5 showed significantly stronger response to a 72-hour median concentration (5 μM) of oxaliplatin treatment (left panel). A stronger response to low and median concentrations of oxaliplatin treatment was also observed in actopaxin siRNA transfected 97L-LUC cells (right panel). (C) Following median oxaliplatin treatment, shRNA cells showed a significantly higher number of cells arrested at G2 phase and decreased number of cells entering G0/G1 phase (left panel). Similarly, siRNA transfected cells also showed an apparently higher number of cells arrested at G2 phase and decreased number of cells entering G0/G1 phase (right panel). (D) When compared with the shCTL cells, shRNA cells displayed a significantly higher proportion of apoptotic cell following 48-hour and 72-hour oxaliplatin treatment at low concentration (left panel). Similarly, actopaxin siRNA transfected cells also displayed a significantly higher proportion of apoptotic cell following 48-hour median concentration of oxaliplatin treatment (right panel). (E) Western blotting demonstrated differential altered expressions of β-catenin/cyclin D1, phospho-mTOR(ser2448)/phospho-4EBP1(thr37/46), p53, and survivin in actopaxin down-regulated cells. (F) Schematic diagram demonstrating how actopaxin repression enhances chemosensitivity to oxaliplatin.

We found that several pathways were involved in the enhanced chemosensitivity to oxaliplatin in actopaxin-repressed cells (Fig. 7E,F). Upon oxaliplatin treatment at low and median concentrations, dose-dependent inhibitory effects on β-catenin and mTOR pathways were observed, as reflected by a dose-dependent decrease in β-catenin/cyclin-D1 and phospho-mTOR(ser2448)/phospho-4EBP1(thr37/46) expression, respectively. Such an inhibitory effect was stronger in both actopaxin shRNA and siRNA transfectants. Expression of the apoptotic protein p53 was induced by oxaliplatin treatment, and the expression was much stronger in actopaxin shRNA and siRNA transfectants. On the other hand, expression of the antiapoptotic protein survivin was induced upon oxaliplatin treatment; however, such induction was attenuated in actopaxin shRNA and siRNA transfectants.

Discussion

This study demonstrates for the first time the clinical significance of actopaxin in HCC, and that actopaxin was involved in the regulation of cell invasion, migration, and metastasis of HCC. While this article was in preparation, Pignatelli et al.[13] showed that actopaxin phosphorylation is required for matrix degradation and cancer cell invasion in their osteosarcoma and breast cancer cell line models, via regulation of RhoGTPase signaling. Their results, in combination with our findings in this study, revealed the significant role of actopaxin in cancer cell metastasis.

In the in vivo study of the effect of actopaxin repression on tumor cell proliferation, to take into account the potential importance of the local environment on the development and progression of cancer, we applied two mouse models. Both the subcutaneous and orthotopic models revealed that down-regulation of actopaxin impaired HCC metastasis. In our orthotopic injection model, all mice developed lung metastasis only, which is concordant with the clinical observation that lung metastasis is the most common extrahepatic metastasis in HCC patients. This could be explained by hematogenous dissemination to the pulmonary capillary network. Interestingly, in our subcutaneous injection model, mice developed liver, lung, and brain metastases. We try to explain such observations by looking at the difference in metastatic mechanism between these two models. We postulate that the tumor microenvironment may play important roles in regulating the development and metastasis of tumor cells. Tumor cells that interact with the microenvironment of the liver may be more favorable in developing lung metastasis, while subcutaneously injected cells may produce a comparably nonspecific metastasis pattern. Two of the mice orthotopically injected with shRNA cells still showed weak luciferase signal in their lungs, suggesting that the orthotopically injected cells were more metastatic to the lung than subcutaneously injected cells. Since orthotopically injected cells were implanted in their native environment, we believed that the orthotopic injection model has a higher clinical value in mimicking the effect of actopaxin repression in the body.

Several studies demonstrated the important roles of β1-integrin in HCC cell migration, invasion, and metastasis.[3, 4] Moreover, the activity of β1-integrin has been found crucial to its functional roles. For example, activation of β1-integrin through its phosphorylation at thr788/789 was only observed in invasive cells.[4] Activation of β1-integrin can be achieved by extracellular agonists (such as transforming growth factor-β1 and autocrine motility factor/phosphohexose isomerase in HCC[3, 4]) as well as intracellular signal transduction molecules, yet to our knowledge little is known about how such an inside-out mechanism is involved in HCC metastasis. This study showed that CD29 is an upstream regulator of actopaxin, whereas suppression of actopaxin attenuated the CD29 activation, suggesting that high CD29 expression correlates with higher expression of actopaxin, whereas high actopaxin expression induces CD29 activity and thus results in HCC cell migration, invasion, and metastasis.

Another important finding in this study is the enhanced chemosensitivity of HCC actopaxin shRNA and siRNA transfectants. Oxaliplatin is a third-generation platinum-derived chemotherapy agent, which, alone or combined with other agents, has been used for advanced HCC in some clinical trials. Oxaliplatin potently inhibits growth and induces tumor cell death in HCC in vitro.[19] Recently, Xiong et al.[20] reported that residual HCC cells after oxaliplatin treatment had increased metastatic potential in a nude mouse model, thus it is reasonable to hypothesize that actopaxin overexpression in HCC cells may be responsible for such an increased metastatic potential of HCC cells. High actopaxin-expressing cells, which are less chemosensitive, can survive oxaliplatin treatment and are enriched in the oxaliplatin-treated population, resulting in enhanced metastatic capacity due to overexpression of actopaxin. Such hypotheses might also be applicable to human HCC patients.

Moreover, this study demonstrates that the enhanced efficacy of oxaliplatin on actopaxin-down-regulated cells was accompanied by stronger inhibition of the β-catenin and mTOR pathways, which are critical cell-signaling pathways that perpetuate the carcinogenic process in HCC, and are thus of interest from a therapeutic perspective, as targeting these pathways may help to reverse, delay, or prevent hepatocarcinogenesis.[21] A stronger induction in p53 expression was also detected in actopaxin-repressed cells. Although the cell line used in this study (MHCC97L) harbored mutant p53, several p53 mutants with a low level of transcriptional activation function have also been shown to induce apoptosis in cancer cells, demonstrating the nontranscriptional apoptotic ability of mutant p53.[22, 23] More important, a recent report demonstrated that the mutant p53 in MHCC97L cells was functional in inducing apoptosis.[24] Furthermore, a weaker elevation of survivin level upon oxaliplatin treatment was observed in actopaxin-down-regulated cells. Survivin, which is selectively expressed in the most common human cancers and appears to be involved in tumor cell resistance against some anticancer agents, has been implicated in the control of cell division and inhibition of apoptosis.[25] In this study, survivin expression was induced by oxaliplatin treatment, indicating that the cells were protected from apoptosis. Although survivin induction was detected regardless of actopaxin expression, such induction was weaker in actopaxin-down-regulated cells, suggesting that the protective effect was reduced after actopaxin down-regulation.

Our findings support that actopaxin might be used as a biomarker for monitoring HCC development and progression. In addition, as actopaxin down-regulation can reduce the metastatic potential of HCC cells and sensitizes them to chemotherapy, it is a potential therapeutic strategy in treating advanced HCC.

Ancillary