Potential conflict of interest: Nothing to report.
Supported by a Direct Grant for Research from the Chinese University of Hong Kong (2009.1.005) and a Collaborative Research Fund from the Hong Kong Research Grants Council (CUHK4/CRF/08).
Genomic amplification of regional chromosome 8q24 is a common event in human cancers. In hepatocellular carcinoma (HCC), a highly aggressive malignancy that is rapidly fatal, recurrent 8q24 gains can be detected in >50% of cases. In this study, attempts to resolve the 8q24 region by way of array comparative genomic hybridization for affected genes in HCC revealed distinctive gains of block of proliferation 1 (BOP1). Gene expression evaluation in an independent cohort of primary HCC (n = 65) revealed frequent BOP1 up-regulation in tumors compared with adjacent nontumoral liver (84.6%; P < 0.0001). Significant associations could also be drawn between increased expressions of BOP1 and advance HCC staging (P = 0.004), microvascular invasion (P = 0.006), and shorter disease-free survival of patients (P = 0.02). Examination of expression of C-MYC, a well-known oncogene located in proximity to BOP1, in the same series of primary HCC cases did not suggest strong clinicopathologic associations. Functional investigations by small interfering RNA–mediated suppression of BOP1 in HCC cell lines indicated significant inhibition on cell invasion (P < 0.005) and migration (P < 0.05). Overexpression of BOP1 in the immortalized hepatocyte cell line L02 showed increase cellular invasiveness and cell migratory rate (P < 0.0001). In both gene knockdown and ectopic expression assays, BOP1 did not exert an effect on cell viability and proliferation. Evident regression of the epithelial-mesenchymal transition (EMT) phenotype was readily identified in BOP1 knockdown cells, whereas up-regulation of epithelial markers (E-cadherin, cytokeratin 18, and γ-catenin) and down-regulation of mesenchymal markers (fibronectin and vimentin) were seen. A corresponding augmentation of EMT was indicated from the ectopic expression of BOP1 in L02. In addition, BOP1 could stimulate actin stress fiber assembly and RhoA activation. Conclusion: Our findings underline an important role for BOP1 in HCC invasiveness and metastasis potentials through inducing EMT and promoting actin cytoskeleton remodeling. (HEPATOLOGY 2011;)
Hepatocellular carcinoma (HCC) accounts for 85%-90% of primary liver cancers and is one of the most common malignancies worldwide.1 Because of its late clinical presentation, patients diagnosed with HCC are often not amenable to curative treatments such as surgical resection and transplantation.2 HCC is hence associated with a high mortality rate. Systemic chemotherapy is the mainstay treatment option for patients with inoperable disease. However, HCC is also highly refractory to most conventional therapeutic modalities. Consequently, the prognosis of HCC patients is extremely poor, with only 5% of patients surviving more than 5 years.3 Given the frequency of intrahepatic and extrahepatic metastases and the impact this has on survival, it is important to understand the molecular characteristics and related biological mechanisms in tumor cell dissemination.
Gains of the chromosome 8q long arm are frequently detected in human cancers, including HCC.4-6 Early molecular investigations on different cancer types, albeit low-resolution, have defined a critical overlapping region to chromosome 8q24.6 Multiple independent studies further emphasized on the importance of this subchromosomal region in cancer development and progression.7-9 Our group and others have reported on recurrent 8q24 gains in 48%-77% of HCC cases.10-12 Recent informatic analysis conducted on the genomic data of >150 HCC tumors further enabled us to underpin 8q24 gain as a major contributory event in the development of HCC.13 Using matrix-based comparative genomic hybridization (CGH) of BAC array analysis on a cohort of >65 HCC samples, German researchers were able to map more precisely the affected genomic loci in HCC, and refined the critical interstitial region of 8q24 to 8q24.23-24.3, which spanned ≈6.82 Mb.14
In this study, we resolved the 8q24.23-24.3 region for affect genes in HCC by way of high-resolution array CGH (aCGH). Distinct gains of the block of proliferation 1 (BOP1) gene at 8q24.3 prompted our further investigations for its role in liver tumorigenesis. We showed that frequent overexpression of BOP1 in primary HCC tumors correlates with advance pathologic features and shorter survival of patients. The prognostic value was also examined in comparison with the proto-oncogene C-MYC (located at 8q24.21), which is also commonly up-regulated in HCC.15, 16 Our result on HCC suggested BOP1 up-regulation was more frequent than C-MYC up-regulation, and may represent an independent prognostic biomarker. Our study also revealed key functions of BOP1 in promoting epithelial-mesenchymal transition (EMT) and invasiveness of HCC cells, which may have implications in metastasis of HCC.
Tumorous and adjacent nontumorous liver tissues were collected from 65 patients (median age 59; 88% male) who underwent curative surgery for HCC at Prince of Wales Hospital, Hong Kong. Informed consent was obtained from each patient, and the study protocol was approved by the Clinical Research Ethics Committee of the Chinese University of Hong Kong. A diagnosis of HCC was confirmed on histologic examination. Patients were predominantly chronic hepatitis B virus carriers (91%), with identifiable cirrhosis in the nontumorous liver indicated in 66% of cases. Tumor stage was classified according to American Joint Committee on Cancer criteria,17 which graded 39 patients as stage T1, 12 patients as stage T2, and 14 patients as stage T3.
Eight HCC samples that displayed 8q24 gains from conventional CGH analysis were subjected to human 44K CGH microarray analysis (Agilent Technologies, Santa Clara, CA) as described.18 The specimens included cell lines (HKCI-1, HKCI-2, HKCI-9, Hep3B, Huh7, and SK-HEP-1)19, 20 and two primary tumors. Briefly, one microgram of genomic DNA was labeled using the Agilent Genomic DNA Labeling Kit PLUS. Hybridized array was scanned by the Agilent Microarray Scanner and signal intensities were extrapolated by Agilent Feature Extraction 9.1. Data computation and normalization was performed using CGH Analytics vesion 3.4. Putative copy number gain was defined by intervals of two or more adjacent probes with log2 ratios >0.8 compared with the log2 ratios of adjoining probes.
Quantitative Real-Time Polymerase Chain Reaction.
First-strand complementary DNA (cDNA) was prepared from HCC tumors, adjacent nontumoral livers and cell lines by random hexamer (Applied Biosystems, Foster City, CA). Real-time polymerase chain reaction (RT-PCR) was performed using gene specific TaqMan Gene Expression Assays (BOP1, Hs00374884_m1; C-MYC, Hs00905030_m1; Applied Biosystems). The Ct values obtained from triplicate reactions were normalized against the endogenous reference 18S RNA (i.e., ΔCt), and fold change relative to the mean value obtained from six normal liver controls was determined using the 2−ΔΔCt method.21 A relative gene expression of ≥2-fold was considered up-regulation.
The HKCI-9 cell line was maintained in AIM V medium supplemented with 10% fetal bovine serum,20 whereas Hep3B22 and L0223 were maintained in DMEM with 10% fetal bovine serum. In BOP1 knockdown experiments, small interfering RNA (siRNA) against BOP1 (LU-014065-01) or si-Control (D001818-10) (Dharmacon Research, Lafayette, CO) was transfected into Hep3B or HKCI-9 cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). In BOP1 overexpression experiments, L02 was transfected with pcDNA3.1-BOP1 or pcDNA3.1 vector using FuGENE HD (Roche Diagnostics, Basel, Switzerland). In validation experiment of RhoA as downstream effector of BOP1, siRNA against RhoA (LU-003860-00) or si-Control (Dharmacon Research) was transfected into pcDNA3.1-vector or pcDNA3.1-BOP1 expressing L02 cells using Lipofectamine 2000.
Cell Viability and Proliferation Assays.
The effect of BOP1 on cell viability was examined by way of 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl tetrazolium bromide (MTT) assay. Cell viability was expressed as a percentage of maximum absorbance of purple formazan at 570 nm from five replicates in three independent experiments. Transfected cells were also examined for cell proliferation by way of 5-bromo-2-deoxyuridine (BrdU) assay. Cells stained with mouse anti-BrdU monoclonal antibody (Roche Applied Science) were measured for signal incorporation at 450 nm. The result was expressed as a percentage of maximum absorbance based on five replicates in three independent experiments.
Invasion and Migration Assays.
The cell motility and cell invasive abilities were assessed by way of Transwell (Corning Life Sciences, Acton, MA) and Matrigel invasion (BD Biosciences, San Jose, CA), respectively. For Transwell migration, 1 × 104 transfected cells (Hep3B, HKCI-9 and L02) were seeded, whereas 8 × 104 (Hep3B and HKCI-9) and 2 × 104 (L02) transfected cells were seeded for the invasion assay. Cells migrated to the underside of the membrane were fixed and stained in hematoxylin or 0.1% crystal violet. Cells were enumerated from >10 microscope fields, and the mean value was expressed as a percentage relative to mock or vector control. Each experiment was performed in replicate inserts, and the mean value was expressed from three independent experiments.
Western Blot Analysis.
Protein lysates were resolved on 8% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred electrophoretically onto polyvinylidene fluoride membrane. After blocking for nonspecific binding, blots were incubated with specified antibodies against BOP1 (1:100 [Ascenion, Munich, Germany]); γ-catenin, E-cadherin, and RhoA (1:1,000 [Cell Signaling Technology, Danvers, MA]); fibronectin (1:1,000 [Sigma-Aldrich, St. Louis, MO]); vimentin (1:500 [Santa Cruz Biotechnology, Santa Cruz, CA]); cytokeratin 18 (1:1,000 [DAKO, Carpinteria, CA]); or glyceraldehyde-3-phosphate dehydrogenase (1:25,000 [Chemicon, Temecula, CA]). After incubation with secondary conjugated to horseradish peroxidase, protein bands were visualized using enhanced chemiluminescence detection (GE Healthcare, Piscataway, NJ).
Cells grown on glass coverslips were fixed in 4% paraformaldehyde and permeabilized using 0.5% Triton X-100. Primary antibody against cytokeratin 18 (1:500 [DAKO]) were visualized using secondary antibody conjugated to Alexa-Fluor 488 (1:250 [Invitrogen]). Filamentous actins were stained with TRITC-labeled Phalloidin (Sigma-Aldrich). Counterstained with DAPI, cells mounted in Vectashield antifade (Vector Laboratories, Burlingame, CA) were examined by confocal microscopy (LSM5 PASCAL, Carl Zeiss).
RhoA Activation Assay.
Hep3B cells transfected with si-BOP1 or si-Control, and L02 cells transfected with pcDNA3.1 or pcDNA3.1-BOP1 were lysed in buffer containing 10% glycerol, 10 μg/mL leupeptin, and aprotinin (Rho Activation Assay Kit, Millipore, MA). Clear cell lysates were added to Rho Assay Reagent (Rhotekin Rho Binding Domain), and agarose beads were collected by centrifugation at 14,000g. After removing the supernatant, beads were resuspended in protein sampling buffer and boiled for 5 minutes. SDS-PAGE and Western blotting was performed and the amount of guanosine triphosphate (GTP)-bound form of RhoA (1:1,000 [Cell Signaling Technology]) were compared with its endogenous level in whole cell lysate.
Expression of BOP1 in HCC tumors and paired adjacent nontumoral livers and functional effects on cell invasion and migration were compared using the paired Student t test. The Mann-Whitney test was used to examine the BOP1 messenger RNA (mRNA) expression with clinicopathological features, including tumor staging and presence of microvascular invasions. The overall and disease-free survivals were calculated from the date of tumor resection to the time of first recurrence (disease-free survival) or death (overall survival), each to a maximum of 120 months. Patients who were lost to follow-up or died from causes unrelated to HCC were considered censored events. Kaplan-Meier survival curves were constructed, and the differences between groups were analyzed using a log rank test. P < 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism 3.02 (GraphPad Software Inc., San Diego, CA). Correlations with patient survival were also investigated by way of Cox regression using MedCalc (MedCalc Software, Mariakerke, Belgium).
BOP1 Is Frequently Up-Regulated in HCC.
Six cell lines and two primary HCC tumors were subjected to aCGH analysis. An example of an aCGH map on 8q24 is shown in Fig. 1A. Among the candidate genes that showed copy number gain, consistent gains of BOP1 were suggested in all eight samples analyzed. Moreover, high-level gains of BOP1 could be detected in three of eight samples. Overall gain of BOP1 ranked as the highest scored event (Fig. 1B). Validation by way of quantitative RT-PCR confirmed up-regulation of BOP1 in the six cell lines studied (Fig. 1C). This corresponded to an increased protein expression of BOP1 as demonstrated on Western blot analysis (Fig. 1D). We also determined that an increased BOP1 protein could be found in primary HCC tumors relative to adjacent nontumoral liver (Fig. 1D). The mRNA level of BOP1 was further examined in a cohort of 65 paired HCC tumors and their matching nontumoral liver. Common up-regulation of BOP1 was identified in 55/65 (84.6%) tumors compared with adjacent nontumoral liver (paired t test; P < 0.0001) (Fig. 2A).
Prognostic Value of BOP1.
Correlative analysis of BOP1 mRNA levels with clinicopathologic features suggested significant association between increased BOP1 expression and the histologic presence of microvascular invasion (P = 0.0059) (Fig. 2B) and advanced tumor stage (T3, P = 0.0042) (Fig. 2C). Up-regulation of BOP1 was also found to predict shorter overall (P = 0.0091) and disease-free (P = 0.0198) survival of patients (Fig. 2D,E). Multivariate survival analysis using the Cox proportional hazards model further indicated that BOP1 up-regulation correlated with a higher hazard ratio (HR) and worse clinical outcomes (overall survival, P = 0.009, HR 3.075; for disease free-survival, P = 0.025, HR 2.160) (Table 1).
Table 1. Multivariate Cox Regression Analysis of BOP1 and C-MYC Expression in HCC
HR (95% CI)
HR (95% CI)
Abbreviations: CI, confidence interval; HBsAg, hepatitis B surface antigen; HR, hazard ratio. Boldface highlights the statistical significance.
To further establish the prognostic value of BOP1, we simultaneously examined the expression of C-MYC, a recognized oncogene generally assumed to be a target gene from the 8q24 region, in the same series of specimens that were subjected to BOP1 evaluations. Limited by RNA availability, only 57 paired tumor and adjacent nontumoral liver were studied. Increased expressions of C-MYC were suggested in 47.4% cases; however, unlike BOP1, significant correlations could not be established with clinicopathologic features (Supporting Fig. 1) and the survival of patients (Fig. 2F,G). In addition, Cox regression analysis did not suggest correlation of C-MYC expression with the HR and clinical outcomes (overall survival, P = 0.681, HR 1.202; disease free-survival, P = 0.754, HR 1.123) (Table 1). Collectively, these data implicate BOP1 an independent prognostic biomarker in HCC.
Functional Role of BOP1 in Cell Migration and Invasion.
In functional examinations of BOP1, siRNA-mediated suppression of BOP1 in HKCI-9 and Hep3B (≈80% knockdown from quantitative RT-PCR) did not suggest an effect on cell viability from MTT assay, nor cell proliferation from BrdU incorporation (Figs. 3A,C and 4A,C). However, marked inhibitory effects on cell invasion through Matrigel and transwell motility were readily observed. Depletion of BOP1 in both Hep3B and HKCI-9 showed significant reductions in the number of cells invaded (P = 0.0021 and P = 0.0009, respectively) and migrated (P = 0.0475 and P = 0.0327, respectively) (Figs. 3D,E and 4D,E). To further support a role for BOP1 in promoting cell motility, ectopic expression of pcDNA3.1-BOP1 in the immortalized hepatocyte cell line L02 showed marked stimulation on cell invasion and migration, where a >50% increase in the percentage of invaded and migrated cells was detected (P < 0.0001) (Fig. 5D,E). In concordance with knockdown study, ectopically expressed BOP1 in L02 cells did not show an effect on both cell viability and proliferation (Fig. 5B,C).
BOP1 Promotes HCC Cell Motility Through EMT Induction.
Studies on BOP1 have proposed a role for this WD40 protein in mediating maturation of large ribosomal subunits in mammalian cells, specifically in the formation of mature 28S and 5.8S ribosomal RNAs.24, 25 To rule out possible involvement of ribosomal biogenesis in the functional phenotypes observed in HCC, we performed ribosomal RNA analysis in BOP1 knockdown cells. In si-BOP1 transfected Hep3B and HKCI-9 cells, Northern blotting did not indicate differences in the maturation of ribosomal proteins compared with si-Control cells (Supporting Fig. 2). Because we were unable to detect noticeable changes in the ribosomal RNA processing in HCC despite perturbed BOP1 activities, it is tempting to speculate that BOP1 has other cancer-related functions that are yet to be determined.
Given that up-regulated BOP1 correlated with enhanced cell migratory abilities of HCC cells, we next examined the EMT as an underlying mechanism. In si-BOP1–treated cells, up-regulation of the cohesive epithelial markers E-cadherin, cytokeratin 18, and γ-catenin was revealed on Western blot analysis. This finding was accompanied by a corresponding down-regulation of the mesenchymal markers fibronectin and vimentin (Fig. 6A). In addition, immunofluorescence staining of cytokeratin 18 showed prominent remodeled organizations in the cytoplasm of si-BOP1–transfected cells compared with cells treated with si-Control (Fig. 6B). Our findings would imply a loss of mesenchymal-like features and reacquisition of epithelial characteristics in BOP1-depleted HCC cells. The involvement of EMT was further supported by the augmentation of mesenchymal markers in BOP1-transfected L02 cells. In BOP1-overexpressing cells, expression of fibronectin and vimentin was elevated above the levels in vector-transfected cells, whereas cytokeratin 18 and γ-catenin were found to be down-regulated (Fig. 6A). Reduced expression of cytokeratin 18 was further verified on immunofluorescence analysis (Fig. 6C). We were able to further substantiate an effect of BOP1 on hepatocyte differentiation in BOP1-transfected and knockdown cells. Quantitative RT-PCR analysis of markers for differentiated hepatocyte function, including albumin, hepatocyte nuclear factor 4a, and P450, showed an increased level of these markers in BOP1-depleted Hep3B and HKCI-9 cells. On the contrary, the expression of these three genes was consistently reduced upon ectopic expression of BOP1 in L02, suggesting a loss of hepatic differentiation phenotype (Supporting Fig. 3).
In promoting the mesenchymal motile phenotype during EMT, circumferential F-actin fibers of the cytoskeleton are replaced by a network of actin fiber polymerization.26, 27 Using phalloidin staining for F-actin, we found greatly reduced F-actin polymerization and stress fiber disassembly in BOP1-depleted Hep3B cells (Fig. 6D), which would signify an effect of actin cytoskeleton remodeling on the motile inhibition observed. The RhoA guanosine triphosphatase (GTPase) is known to play a regulatory role in the F-actin fiber rearrangements, including the formation of stress fibers.28, 29 We therefore examined the endogenous RhoA status under the influence of BOP1. Results obtained from in vitro GTP pull-down assay in Hep3B and L02 concurred to suggest that RhoA is a likely downstream effector of BOP1 (Fig. 7A).
To establish the association between RhoA activation and BOP1 expression, we further examined the effect of RhoA knockdown in the presence of BOP1 expression. In BOP1-expressing L02 cells, siRNA knockdown against RhoA readily reduced L02 cell invasion by 40% (P = 0.006) and migration by 60% (P = 0.021) when compared with si-Control experiments (Fig. 7B). On the other hand, the effect of RhoA knockdown in vector-transfected L02 cells showed behavior similar to that seen in si-Control assays and did not suggest significant changes in cell motility and invasion (P > 0.05). Our findings suggest that the BOP1-mediated cellular invasiveness and motility were dependent on the activation of RhoA pathway.
In this study, we identified BOP1 as a target oncogene within the common overrepresented region 8q24 in HCC. We also established a functional role for this gene as a contributory factor in EMT and invasiveness of HCC cells. The gene BOP1, also known as block of proliferation 1, is a WD40 protein that was first isolated from the cDNA library screening for growth-related sequences in mouse embryonic fibroblasts.30 Studies on mouse fibroblasts had in fact laid much of the earlier groundwork on the role of BOP1, which implicated BOP1 as an integral component of the ribosomal RNA processing machinery.24, 25, 31 Nevertheless, recent publications have begun to reveal the existence of extraribosomal activities of BOP1. In yeast, BOP1 was shown to play in a role in maintaining the genomic integrity through its function in the accurate chromosome segregation and mitosis at the metaphase/anaphase transition.32 Studies on colorectal cancers have indicated common BOP1 overexpression in cases with 8q gain and have demonstrated that deregulation of BOP1 would likely disrupt vital pathways contributory to colorectal tumorigenesis.33, 34 It was also reported in colorectal cancer that gene dosage increase of BOP1 was independent of that of C-MYC and occurred more frequently.34 In line with this finding, our results on HCC also suggest more frequent up-regulation of BOP1 (84.6%) compared with C-MYC (47.4%). Moreover, contrary to the results obtained from C-MYC, we were able to establish significant BOP1 up-regulation in HCC tumors relative to adjacent nontumoral livers and a positive correlation between BOP1 expression and advanced clinicopathologic features. In addition, multivariate Cox regression analysis revealed that BOP1 may represent an independent prognostic biomarker in predicting overall and disease-free survival of HCC patients.
Functional analysis by way of RNA interference–mediated gene suppression of BOP1 in overexpressing HCC cell lines showed marked reductions in cell invasion and migration but not cell proliferation. In concordance with the knockdown studies, ectopic expression of BOP1 in immortalized L02 hepatocytes promoted cell migratory and invasive abilities. The lack of detectable function for ribosomal biogenesis in these phenotypes prompted our investigation of an alternate mechanism in BOP1-mediated cell motility. We found that BOP1 depletion caused rapid regression of EMT features, where a gain in the expression of epithelial cell adhesion molecules such as E-cadherin and reorganization of cytoplasmic cytokeratin 18 were revealed. In BOP1-transfected L02 cells, a striking increase in the expression of mesenchymal markers (vimentin and fibronectin) concurred with enhanced cell motility and invasive phenotypes. Cumulating evidence has indicated that EMT mediates tumor progression including local invasion, dissemination from the primary tumor, intravasation into blood circulation, and metastasis.27, 35 EMT is a complex process requiring extensive changes in cell adhesion and morphology and activation of signaling paths; our findings suggest that BOP1 may represent an upstream molecule that can induce this transition. It is well recognized that RhoA plays an oncogenic role in enhancing cell contractility through actin rearrangements and stress fiber formation.28, 29, 36 We found considerable depolymerization of actin stress fibers upon BOP1 knockdown, which coincided with a reduced RhoA-GTPase activity. Furthermore, RhoA depletion experiments in BOP1-expressing L02 cells showed greatly reduced cell motility and invasive phenotypes, which further supports RhoA as a downstream effector of BOP1. Our findings implicate a role for RhoA-GTPase activation in mediating the functional effects of BOP1.
Proteins containing WD40 repeats can be found in a range of regulatory proteins that cover a wide variety of physiological functions, including signal transduction, RNA processing, and cytoskeletal dynamics.37-39BOP1 is an evolutionary conserved gene that contains seven WD40 repeats in humans. Recent studies have suggested a role for BOP1 in tumorigenesis33, 34 and microtubule dynamics by interacting with end-binding protein 1,40 which is also a well-known binding partner of the adenomatosis polyposis coli tumor suppressor.41 Here, we defined oncogenic functions of BOP1 in promoting HCC cell migratory and invasive activities through the induction of EMT, which is a key regulatory program activated during cancer invasion and metastasis. The possibility that BOP1 may exert an effect on cancer invasion machinery suggests that there are as yet undescribed cancer-related functions of BOP1. The fact that BOP1 is also commonly overexpressed in a majority of HCC tumors holds prognostic significance. Further elucidation of BOP1's capacity in coordinating protein–protein interactions and cellular processes will contribute to therapy designs and inhibition of HCC metastasis.