Main

Extensive genetic characterizations on hepatocellular carcinoma have indicated common copy number gains on chromosomes 1q, 8q, 17q and 20q, and losses on 1p, 4q 8p, 13q, 16q and 17p.1, 2, 3 While these changes have been considered early genetic events in liver carcinogenesis, the molecular events by which specific genomic alterations contribute to the progression of hepatocellular carcinoma remain elusive. Our group has previously conducted large-scale correlative analysis on the genomic data derived from early and advanced stages hepatocellular carcinoma tumors with an aim to define the genetic aberrations involved in disease progression.4 A number of regional copy gains were indicated, among which chromosome 7q34–q36 gains were common in the advanced metastatic hepatocellular carcinoma.4

Distal chromosome 7q gains have been suggested in association with biliary tract carcinoma progression,5 and higher-stage non-small cell lung cancer with nodal involvement.6 Previous studies have also demonstrated a positive association between prostate cancer progression and copy gains of distal chromosome 7q,7 and the amplification of chromosome 7q31–q36 region in the progression of gliomas to grade III or IV.8 Taken together, it is likely that gene(s) residing on chromosome 7q34–q36 may have a role in cancer progression. In the current study, a combined analysis using of high-resolution genomic and transcriptional mapping was performed to identify potential tumor-related gene(s) residing on chromosome 7q34–q36. Using a complementary DNA array platform with an average resolution of ∼91 kb on chromosome 7, mapping studies where performed on a hepatocellular carcinoma tumor that displayed regional chromosome 7q34–q36 high-level gains from metaphase comparative genomic hybridization analysis. This integrated analysis has permitted recognition of candidate genes and minimized the number of genes required for subsequent quantitative reverse transcription-polymerase chain reaction validations. The functional effects of an identified candidate gene on hepatocellular carcinoma cell proliferation and motility were further investigated.

Materials and methods

Patients

Tumorous hepatocellular carcinoma and the corresponding adjacent nonmalignant liver tissues were obtained from 52 patients who underwent curative surgery for hepatocellular carcinoma at the Prince of Wales Hospital, Hong Kong between 1997 and 2002. Histological examination confirmed the diagnosis of hepatocellular carcinoma, and the presence of underlying cirrhosis in 80.8%. Serological analysis indicated chronic viral hepatitis B carriage in 94.0% of patients. The American Joint Cancer Center tumor staging system classified one case as T1, 30 cases as T2, 12 cases as T3 and 9 cases as T4. Statistical analysis did not suggest differences in the clinicopathological characteristics between the early T1/T2 and advanced T3/T4 stage patients (P>0.05), except for the presence of multifocal feature (P<0.0001) (Table 2). Five year follow-up information on 32 cases showed tumor recurrence in 15 patients.

Table 2 Clinicopathological information of 52 hepatocellular carcinoma patients
Full size table

Complementary DNA Microarray-Based Comparative Genomic Hybridization

The complementary DNA microarray chip utilized contained 1738 complementary DNA clones that spanned an average resolution of ∼91 kb along the entire human chromosome 7 (Ontario Cancer Institute, Toronto, Canada). Complementary DNA microarray-based comparative genomic hybridization was performed according to method described in Beheshti et al.9 A hepatocellular carcinoma cell line, HKCI-6, that displayed high-level gains of regional chromosome 7q34–q36 was employed in the mapping analyses. HKCI-6 was established from the tumorous liver tissue of a chronic hepatitis B virus infected male patient, who presented a stage T3 hepatocellular carcinoma tumor with multifocal lesions. DNA from HKCI-6, and a reference pool (lymphocytic DNA from three healthy subjects) were differentially labeled with Cy5-dCTP and Cy3-dCTP using the RadPrime DNA labeling system (Gibco Invitrogen Corporation). Labeled DNA was combined with 30 μg human Cot-1 DNA, 20 μg poly(dA-dT) and 100 μg yeast tRNA in DIGhyb buffer prior to overnight hybridization in a dark chamber at 37°C. Following posthybridization washes in 1 × saline sodium citrate/0.1% sodium dodecyl sulfate at 45°C and gentle rinsing in 1 × saline sodium citrate at room temperature, hybridized signals were captured by the ScanArray 5000 scanner (GSI Lumonics, Billerica, MA) and analyzed by the GenePix Pro 3.0. Data manipulation was facilitated using the Normalization Suite v1.63 (http://www.utoronto.ca/cancyto/). Results from duplicate spots and dye swap experiments were calculated and incorporated with the physical map locations of complementary DNA in sequential order of megabase distances. ‘Test to reference’ ratios greater than twofold were considered as copy number gain.

Expression Microarray

Using complementary DNA microarray chip as the array-comparative genomic hybridization study (Ontario Cancer Institute, Toronto, Canada), expression array analysis was performed according to procedure previously described from our group.10 Ten micrograms of total RNA from HKCI-6 and reference pool (3 normal livers acquired from CloneTech, Stratagene and Ambion) were reverse transcribed by AncT messenger RNA primer using Supescript II reverse transcriptase separately (Gibco Invitrogen Corporation). Complementary DNA differentially labeled with Cy3-dCTP and Cy5-dCTP were combined with 20 μg calf thymus DNA and 100 μg yeast tRNA prior to hybridization onto array slides. Following posthybridization washes, image acquisition and data analysis were as described for complementary DNA microarray-based comparative genomic hybridization. ‘Test to reference’ ratios greater than twofold were considered as upregulations.

Quantitative Reverse Transcription-Polymerase Chain Reaction

Total RNA extracted from HKCI-6 and primary hepatocellular carcinoma tumors using TRIzol reagent (Gibco Invitrogen Corporation) were treated with 1.0 U Rq1 RNase-free DNase (Promega Corporation) prior to reverse transcription. To ensure no genomic DNA carryover, no-reverse transcription polymerase chain reaction was first performed using primer sets for genomic beta-globin gene (sense: 5′-GAA GAG CCA AGG ACA GGT AC-3′ and antisense: 5′-CAA CTT CAT CCA CGT TCA CC-3′). First strand complementary DNA was prepared from total RNA using the random hexanucleotide primer and TaqMan Reverse Transcription Reagent (Applied Biosystems). Primers used for zyxin were (sense) 5′-CAC CGG CTC AGA ACC AAA-3′ (antisense) 5′-TGC TCC ATG TCC TGC ATT AG-3′ and for ATP-binding cassette sub-family F member 2 were (sense) 5′-ACA ACG TGT GCA CCC TGA-3′ (antisense) 5′-CCA GCT GCT CAA ACA CGT C-3′, respectively. Primers for the internal control glyceraldehyde-3-phosphate dehydrogenase were (sense) 5′-ATG GGT GTG AAC CAT GAG AAG-3′ and (antisense) 5′-AGT TGT CAT GGA TGA CCT TGG-3′. The emission intensity from SYBR green binding to double-stranded DNA was detected by the iCycler detection system (BioRad Laboratories, Hercules, CA). Triplicate PCR amplifications were performed for each sample. No template polymerase chain reactions served as negative controls in all experiments. Relative expression levels in primary tumors with reference to normal liver RNA from three individuals were expressed as 2 - Δ C t Expression levels greater than twofold were considered as up-regulation.

Expression Knockdown by RNA Interference

Hep3B was cultured in Dulbecco's modified eagle medium according to recommendation from the American Tissue and Cell Collection (USA). Specific post-transcriptional knockdown of zyxin expression were achieved using the SMART small interfering RNA pool (Dharmacon Research Inc., USA). A siControl pool that does not align to any known mammalian mRNA served as nonspecific small interfering RNA in mock experiments. Control experiment using transfectant only was also performed. Cells at exponential growth were transiently transfected using Lipofectamine 2000 in fetal bovine serum-free medium. After 5 h incubation, transfection mixtures were replaced with growth medium.

Cell Proliferation Assay

Cell viability was assessed using the thiazolyl blue tetrazolium bromide assay. Approximately 6 × 103 cells were seeded in quintuplicate onto 96-well plates at day 0. The assay was performed at 24 h intervals. Following removal of medium, 50 μl of 1.0 mg/ml thiazolyl blue tetrazolium bromide was added to the cells and incubated at 37°C for 4 h. Using SPECTRA Rainbow v2.02 plate reader (TECAN Austria Ges.m.b.H), cell viability was measured by the colorimetric product developed at absorbance 570 nm. Results were averaged from three independent experiments.

TdT-Mediated dUTP Nick End Labeling Assay

Transfected cells were fixed in 4% paraformaldehyde and permeabilized with 0.2%. Triton X-100 prior to fluorescein-12-dUTP incorporation using the TdT-mediated dUTP nick end labeling assay (Promega Corporation). The reaction took place in a dark chamber at 37°C for 1 h. Slides washed in 2 × saline sodium citrate were nuclear stained with 4′,6-diamidino-2-phenylindole. The evaluation of fluorescent signals for apoptotic cells was performed on a Leitz DM RB fluorescence microscope (Leica Microsystems Ltd.).

Cell Migration and Invasion Assays

The cell migration capacity was estimated by the Costar transwell inserts (8 μm pores) (Corning Incorporation) and the cell invasive capacity was estimated using the BD BioCoat Matrigel invasion chambers (8 μm pores, 24-wells) (BD Bioscience). Cells (1 × 104 for migration study and 5 × 104 for invasion assay) were resuspended in serum-free medium and seeded onto the top chamber of prewet inserts. After 48 h incubation, cells on top of the membrane were wiped off and the inserts were methanol fixed, stained with hematoxylin and eosin, and alcohol dehydrated. Numbers of cells invaded to the underside of the membrane were scored from five microscope fields (magnification, × 100). Each condition was performed in triplicates and results described were the average from two independent experiments.

Statistical Analysis

Correlation of gender, cirrhosis, hepatitis B viral status and multifocal presentation between early and advanced stage hepatocellular carcinoma were examined using Pearson's χ2 test. The Student's t-test was performed to compare the differences in age and serum alpha-fetoprotein level. The expression levels of candidate gene(s) between different groups of hepatocellular carcinoma tumors were compared by Student’s t-test and χ2 test. Results derived from cell proliferation assays were examined using one-way analysis of variance. In invasion and migration assays, Student’s t-test was applied to investigate potential effects upon treatments. A difference was considered significant when P-value ≤0.05. All statistical analysis was performed using SPSS for Windows 10.0 software and Graphpad Prism 3.0.

Results

Candidate Genes on Chromosome 7q34–q36

Simultaneous genomic and transcriptional mapping analyses on chromosome 7 were performed on HKCI-6 using the complementary DNA microarray platform. Comparing the profiles derived from metaphase comparative genomic hybridization with the complementary DNA microarray-based comparative genomic hybridization, a concordant copy gain in the region chromosome 7q34–q36 was indicated (Figure 1a). A total of 27 expressed sequence tags along chromosome 7q34–q36 demonstrated copy number gains (microarray ratio ≥2.0) (Table 1a), although parallel expression microarray analysis indicated the upregulations (microarray ratio ≥2.0) of only 12 transcripts from within the same region (Table 1b).

Figure 1
figure 1

Integrative complementary DNA microarray-based comparative genomic hybridization and transcriptional mapping on chromosome 7q34–q36. (a) Left: Metaphase comparative genomic hybridization revealed a regional high-level gain of 7q34–q36 in HKCI-6. Right: Combined complementary DNA microarray-based comparative genomic hybridization# and expression analyses identified two candidate genes, ZYX and ABCF2, with increased copy gain and distinct over-expression. (b) Over-expression of messenger RNA level of ZYX was confirmed by quantitative reverse transcription-polymerase chain reaction* in HKCI-6. (c) Relative expression level of ZYX in hepatocellular carcinoma patients. ZYX expression was found to be higher in cases with multifocal tumor lesions than cases with solitary tumor.

Full size image
Table 1 Copy number gain and upregulated genes on chromosome 7q34–q36. (a) Genes that displayed copy number gains as depicted from array-based comparative genomic hybridization analysis. (b) Upregulated genes identified from expression array analysis. Genes selected for quantitative reverse transcription-polymerase chain reaction validation are in bold
Full size table

When aligning the results from genomic mapping and transcriptional analysis, nine transcripts demonstrated simultaneous DNA copy number gains and messenger RNA upregulations. These included KIAA1277, KIAA0773, zyxin (ZYX), EPH receptor A1 (EPHA1), protein disulfide isomerase family A member 4 (ERP70), cyclin-dependent kinase 5 (CDK5), ATP-binding cassette sub-family F member 2 (ABCF2), mixed-lineage leukemia 3 (MLL3) and ring finger protein 32 (RNF32). In particular, ZYX and ABCF2 displayed distinct over-expressions at 9.3-fold and 17.6-fold, respectively, compared to other candidates (Figure 1a). Further quantitative reverse transcription-polymerase chain reaction verification in the same specimen, however, confirmed the overexpression of only ZYX at 16.4-fold, whereas contrary to array finding a downregulation of ABCF2 was indicated (Figure 1b).

Expression of ZYX mRNA in Primary Hepatocellular Carcinoma Tumors

The ZYX expression was further examined in a cohort of 52 hepatocellular carcinoma tumors and the corresponding adjacent nonmalignant liver tissue (31 early stages T1/T2 and 21 advanced stages T3/T4) (Table 2). Overexpression of ZYX was suggested in ∼33% cases (17/52 cases) with an average fold gain of 83.5±581.0 compared to their nontumorous counterpart. Although correlative analysis did not suggest differential ZYX expressions with tumor staging (P>0.05), ZYX expression was found to be on average 60-fold higher in cases that presented with multiple tumor lesions (n=20) (211.5±936.9-fold) when compared with those presenting with solitary tumors (n=32, 3.5±6.3-fold, Figure 1c). Furthermore, among the 32 patients in whom 5 year follow-up was available, a substantial portion of those with recurrent disease (14/15 cases) had overexpression of ZYX by >10-fold compared to those nonrecurring patients (P=0.05).

Effect of ZYX Knockdown on Hepatocellular Carcinoma Cells

The Hep3B cell line was employed in the functional study of ZYX since Hep3B displayed a motile invasive phenotype and at the same time demonstrated an over-expression of ZYX (21.7-fold from quantitative reverse transcription-polymerase chain reaction analysis). More than 80% reduction in the messenger RNA expression of ZYX was achieved in Hep3B cells by RNA interference (siZYX). Quantitative reverse transcription-polymerase chain reaction also indicated the inhibition of ZYX expression to be maintained for up to 7 days. There was no observable change in cell morphology upon siZYX transfection compared to mock and control experiments, although a reduction in the relative cell proliferation by ∼30% in siZYX-transfected cells was indicated compared to mock experiments (Figure 2a). Analysis of knockdown effect on apoptosis by TdT-mediated dUTP nick end labeling NEL assay, on the other hand, did not reveal differences between siZYX, mock and control experiments (Figure 2b).

Figure 2
figure 2

Effect on cell growth upon siZYX transfection. (a) A reduced growth rate by ∼30% was suggested in cells with suppressed ZYX expression. (b) Apoptotic status in cells transfected with (i) siZYX was found similar to that observed in (ii) mock and (iii) control experiments. Cells undergoing apoptosis were indicated by green fluorescent signals and nuclei were stained blue with 4′,6-diamidino-2-phenylindole.

Full size image

There was a marked difference in the invasion index of siZYX-treated cells and mock at 43.8 and 9.5 relative to control (P=0.005) (Figure 3a). Similarly, in the migration assay, a significant reduction of chemotatic cell migration of siZYX-treated cells was found (P=0.002) (Figure 3b and c). The migration index of siZYX-transfected cells, and mock were 62.1 and 33.4, respectively compared to control experiment.

Figure 3
figure 3

Effect on cell migration and cellular invasion upon ZYX expression knockdown. (a) A significant reduction in cell invasion through the extracellular matrix was observed in siZYX-transfected cells. **P<0.005. (b) In cell migration assay, a significant reduction was also observed in cells with ZYX expression knockdown. Representative images of cells invaded through the matrix gel coating to the underside of migration chamber are shown in (c). NS: not significant, **P<0.005.

Full size image

Discussion

Metastasis remains the major obstacle to curative therapies for patients with hepatocellular carcinoma. Moreover, the high incidence of postoperative recurrence associated suggests the presence of pre-existing micro-metastasis prior to surgery.11 This observation underscores the need to understand the molecular characteristics and related biological mechanisms that lead to tumor cell dissemination. In this study, we have attempted to map the chromosome 7q34–q36 region for potential candidate(s) that may be involved in hepatocellular carcinoma progression. Integrative positional mapping by array-based comparative genomic hybridization and expression array has enabled us to identify a single candidate gene, ZYX. In order to confirm upregulation of ZYX, its expression was examined in a cohort of hepatocellular carcinoma tumors among which ZYX over-expressions occurred in around 33% of cases when compared to the adjacent cirrhotic or fibrotic liver tissues. Overexpression of ZYX has also been demonstrated in melanoma, compared to its biological precursor melanocytes.12 However, although a relationship between ZYX upregulation and progression of bladder cancer was suggested,13 our correlative analysis did not demonstrate significant association between messenger RNA levels of ZYX and advanced stage hepatocellular carcinoma tumors. Rather, expression levels of ZYX in patients with multiple tumor lesions were elevated when compared to cases with solitary tumor (∼60-fold higher than solitary hepatocellular carcinoma). As many studies have shown that intrahepatic satellite nodules in hepatocellular carcinoma are mostly metastases of the primary lesion,14, 15 our finding may be interpreted as indicating a biological role for ZYX in hepatocellular carcinoma tumor cell disseminations. We have also shown ZYX overexpressions to be more common in those patients who developed recurrences following curative hepatectomy.

ZYX is a zinc-binding phosphoprotein (∼61 kd) that is characterized by three LIM domains. It is abundant at focal adhesions and along the actin cytoskeleton.16 Our data suggest that ZYX suppression inhibits cell growth. Consistent with this observation ZYX expression has been shown to correlate directly with cell proliferation and in decreased the doubling time in melanoma cells.12 As there was no difference in the apoptotic status of ZYX-deficient, mock and control cells, the effect of ZYX on hepatocellular carcinoma cell growth is unlikely to be mediated by antiapoptotic mechanisms. The ability of ZYX to enhance tumor growth may be attributable to the LIM domains, which have previously been shown to regulate cell proliferation and differentiation.17 We also examined the role of ZYX in hepatocellular carcinoma cell invasion and migration, since ZYX has a reported role in mediating cell motility in vitro.18 Following ZYX expression knockdown, the ability of hepatocellular carcinoma cells to pass through the extracellular matrix and to migrate towards the chemo-attractant fetal bovine serum were both attenuated. These results suggest that ZYX expression plays a role in hepatocellular carcinoma cell motility and potentially underlined its function in tumor cell dissemination.

There is evidence to suggest that ZYX is recruited from cytoplasm to integrin-mediated focal complexes where it transforms the focal complexes into a focal adhesion.19 It is therefore possible that ZYX affects hepatocellular carcinoma cell migration through its interaction with integrins at sites of focal adhesions. Furthermore, it has been reported that ZYX is rapidly mobilized from the focal adhesions to actin stress fibers where it can mediate cell migration through actin tractions.18, 20 Taken together, our data suggest that ZYX may have enhanced hepatocellular carcinoma cell migration and intravasation through its action on actin cytoskeleton and promoted cell dissemination as a part of the integrin-signaling pathways.

In the current study, our integrated mapping analyses have demonstrated the efficacy of targeting candidate genes within aberrant genomic loci. The characterization of ZYX-deficient hepatocellular carcinoma cell has led to the recognition of its potential role in cell motility, invasion and proliferation. Further investigations into the molecular pathways of ZYX and its interaction with actin cytoskeleton should shed light on the metastatic behavior of hepatocellular carcinoma.