| Background and aimsHepatocellular carcinoma (HCC) accounts for a high proportion of liver cancer and ranks third in anniversary cancer mortality rates worldwide due to its late detection and fast progression. Although the risk factors such as hepatitis B or C viral infection, pre-existing liver cirrhosis, excessive alcohol consumption, and consumption of aflatoxin B1contaminated food are well known, the molecular mechanism in hepatocarcinogenesis has not been fully elucidated. HCC development is a multistep process involved activation of oncogenes and silencing of tumor suppressor genes.We have previously performed a2-Mb array based CGH analysis of25-paired HCC DNA specimens derived from both timorous and their corresponding non-cancerous tissues. Based on our aCGH data, losse of chromosome13q21.32where PCDH9gene mapped was found in6tumor specimens (24%) and gains in1(4%) of cases. Protocadherin9(PCDH9,13q21.32) gene in hepatocellular cancer (HCC) tissues. PCDH9, a member of the cadherin superfamily, was implied to act as a tumor suppressor in glioma. But its functions and intracellular signal transduction are largely unknown, especially in HCC. Here we analyzed its epigenetic inactivation, biological effects, and prognostic significance in HCC.Cadherin superfamily, a large family of trans-membrane or membrane associations glycoprotein, plays an essential role in regulation of organ and tissue development during embryogenesis and in formation of stable cell-cell junctions and maintenance of normal tissue structure in adult organisms. Over the last decade, several members of the protocadherins subfamily including PCDH8, PCDH10, PCDH17and PCDH20have been identified and characterized as tumor suppressor or as pro-oncogenic candidates. PCDH9is a member of the81-protocadherin subfamily and predominantly expressed in the nervous system. The PCDH9gene is located in the region of chromosome13q21.32and encodes a protein with molecular weight of approximately18kDs. Allelic losses of13q is one of the most frequently genetic alteration found in HCC and in other cancers, implied that13q region likely contain putative tumor suppressor genes. Recent studies have indicated that PCDH9was down modulated in non-nodal mantle cell lymphoma (MCL) and glioblastoma, which suggested that the PCDH9may be a potential tumor suppressor gene. However, the role of PCDH9to tumor development and growth has not been fully elucidated. Here we analyzed its expression,epigenetic inactivation, biological effects, and prognostic significance in HCC.Materials and methodA total of120HCC patients who underwent surgery in Henan Tumor Hospital (Zhengzhou, Henan, China) from2009to2013were enrolled in this study (ranging age from11years to80years, mean age=53.4±11.0years; male:female=77:43)。All patients were serum HBsAg or HBV-DNA positive;102of them had accompanying liver cirrhosis. All HCC diagnoses were confirmed pathologically and their tumor stage was determined according to the2002International Union Against Cancer TNM Classification System.Human liver cell lines (SNU-182, SNU-387,SNU-449, SMMC-7721,, PLC/PRF5, SK-HEP-1, Huh-1, Huh-7, Hep3B) were stored in our laboratory and maintained in RPMI1640or Dulbecco’s modified Eagle’s medium (DMEM) with10%fetal bovine serum (FBS). All incubations were performed at37℃and5%CO2.The expression of PCDH9was verified by real-time quantitative RT-PCR, Western blot and immunohistochemistry assays. The promoter methylation status was analysed by combined DNA methylation-sensitive and methylation-dependent restriction endonuclease digestion, followed by subsequential quantitative PCR assay.. The biological function of PCDH9in HCC cell lines was investigated using various cell assays.All statistical analyses were performed with SPSS13.0software, all the results are expressed as mean±SD. Paired t-test statistical analysis (two-tailed, P<0.05marked a significant difference) was used to determine the difference of PCDH9mRNA expression. The comparison of patient’s characteristics and methylation states was analyzed by chi-square test (χ2test). MTT assay using factorial analysis of variance design data. The differences between the two groups were determined by independent sample t-test. ALL were2-sided statistical tests and a P-value of less than0.05was taken as statistically significantly difference.Results:1. Down-regulation of PCDH9expression in HCC tissuesReal-time RT-PCR demonstrated that compared to normal liver tissues, the down-regulated of PCDH9mRNA was found in five (Huh-7, SMMC-7721, SK-HEP-1, SNU-182and SNU-449) of8HCC cell lines.. In order to further detect the PCDH9level in the primary human HCC tissues,120paired HCC tumor tissues and their adjacent non-tumor tissues analyzed. In general, PCDH9mRNA level in tumor tissues was significantly lower than that in corresponding non-tumor tissues (P=0.032). Up to61%(73/120) of tumor tissues showed significant decrease of PCDH9expression (2-ΔCT<0.5). This tendency was further confirmed by western blot and immunohistochemistry assays. Moreover, statistical analysis showed that the PCDH9low expression was significantly correlated with the malignant portal vein invasion (65.8%vs.34.2%, P=0.019). Comparison of survival curves was done with the log-rank (Mantel-Cox) test, percentage of survival in PCDH9expression T<N group was lower than PCDH9expression T≥N group, although this difference did not reach statistical significance (P=0.258), the tendency suggest that PCDH9is a candidate tumor suppressor gene.2.Allelic loss and methylation mediate the down regulation of PCDH9expression in HCC tissuesBy aCGH assay, aberrant copy number variation of13q21.32locus where PCDH9located in7of25tumor tissues tested, including6deletions (24%) and1gain. Further sequencing analyses confirmed the LOH of PCDH9in these aCGH-arrayed HCC tissues. Meanwhile, to test if hypermethylation also contributed to its down-regulation in HCC, the PCDH9promoter methylation intensity (MI) in the HCC cell lines and primary tumor and their non-tumor tissues were tested. Hypermethylation (HM>10%) was found in SMMC-7721cells and in24of111(about22%) paired tissues successfully measured. Consistently, treatment with DNA demethylation reagent5’-Aza restored PCDH9gene expression in SMMC-7721cells, but not in the PCDH9promoter un-methylated cell lines (PLC/PRF/5, Huh-7, SK-HEP-1, SNU-182, SNU-449, Hep3B and SNU-387). What’s more, the expression of PCDH9in methylation group was significantly lower than that in un-methylation group (P=0.035). Strikingly, statistical analysis revealed a significantly association of PCDH9hypermethylation with larger tumor size (≥5cm; P=0.044) and worse intrahepatic dissemination (P=0.020).3. PCDH9inhibits tumor cell proliferation and xenographt tumor formationTo explore the role of PCDH9in HCC, HCC cell line SNU-449with extremely low endogenous PCDH9level were transfected with either PCDH9expression construct or pIRES2-EGFP vector control. Two HCC cell lines (Hep3B and PLC/PRF/5) with high endogenous PCDH9level were transfected with either PCDH9knockdown plasmid (shRNAl and shRNA2) or pAAV-U6vector control. After4weeks of G418constant selection, we found that in SNU-449cells, PCDH9-transfected cells formed notably less colonies in number and smaller in size, compared with their empty vector transfectants. While in Hep3B and PLC/PRF/5cells compared with their empty vector transfectants, shRNAl and shRNA2formed more colonies in number and larger in size. Real-time RT-PCR showed that in ectopic PCDH9-transfected SNU-449cells, the PCDH9mRNA level46.8-fold increase compared with the vector-control cells. Meanwhile in Hep3B cells, shRNAl and shRNA2showed6.9-fold and4.9-fold decrease compared with their empty vector transfectants. Similarly, in PLC/PRF/5cells, shRNA1and shRNA2showed5.1-fold and3.8-fold decrease compared with their empty vector transfectants. Western blot confirmed result of the real-time RT-PCR.The cell proliferation capacity of these ectopic PCDH9stably expressing or knockdown exogenous PCDH9was analyzed. The MTT assay showed that ectopic PCDH9stably expressing SNU-449cells grew significantly slower than their controls (p<0.001). Consistently, flow cytometry analysis revealed the increase of cells in G1phase and decreased in S phase, but no increase in sub-G1phase cells was observed in SNU-449cells compared to their controls, respectively. These data suggested that PCDH9inhibited HCC cells proliferation via inducing G1phase arrest, but not inducing apoptosis. Furthermore, we detected the expression of cell cycle progression regulators cyclin E, p21waf1/cip1and p27by western blot. We found that in SNU-449cells, PCDH9expression significantly increased the protein level of p21waf1/cip1and p27, but suppressed cyclin E expression, as compared with the control cells, which suggested that PCDH9might inhibit cell proliferation by regulating the expression of cell cycle regulator.In contract, the MTT assay showed that it was significantly increased in the PCDH9knockdown Hep3B and PLC/PRF/5cell line (shRNAl and shRNA2). Further anchorage independent growth assay showed that in SNU-449cells, PCDH9-reexpressed cells formed much fewer and smaller colonies in soft agar than the control counterparts(P<0.001). Conversely, in Hep3B cells, PCDH9knockdown (shRNAl and shRNA2) induced cells formed more and bigger colonies in soft agar than that of controls(P<0.001). Similar results showed in PLC/PRF/5cells. This revealed that cells stably overexpressed PCDH9had decreased colony-forming efficiency relative to vector-transfected controls, whereas the cells that knockdown PCDH9had greater colony-forming efficiency.In addition, Xerograft tumor formation assay was carried out using nude mice. Four days after injected with PCDH9re-expressed SNU-449cells, tumor appeared in3of the6mice. By contrast, tumor appeared in5of6mice injected with control SNU-449cells. Moreover,5weeks after injection, significant differences in average tumor size were observed:545.8±131.9mm3of vector control vs.171.5±92.10mm3of PCDH9re-expressed SNU-449cells. On the contrary, Four days after injected with knockdown PCDH9(shRNAl and shRNA2) Hep3B cells, tumor appeared in5of the6mice. By contrast, tumor appeared in4of6mice injected with control Hep3B cells. Moreover,5weeks after injection, significant differences in average tumor size were observed:220.3±60.8mm3of vector control vs.372.7±49.8mm3of shRNAland309.2±38.6mm3of shRNA2(P<0.001). Furthermore, the similar results showed in PLC/PRF/5cells. These data indicated that PCDH9could inhibit HCC cell proliferation and reduce their malignant transformation capacity in vitro and in vivo.4. PCDH9suppresses HCC cells migrationSince statistic analysis revealed a strongly association of methylation mediated PCDH9expression silencing with HCC intrahepatic dissemination, we performed wound-healing assay and transwell migration assay to assess the effect of PCDH9on HCC cell motility and migration. Our data showed that the spread of PCDH9-transfected SNU-449cells along the wound edges were remarkably slower than control cells at24h and36h post scratching (P<0.001). On the contrary, in Hep3B cells, knockdown PCDH9(shRNAl and shRNA2) could promote HCC cell migration at24h and36h post scratching compared with the vector control. Meanwhile, the similar results showed in PLC/PRF/5cells.Furthermore, the effect of ectopic overexpression or knockdown PCDH9on HCC cell migration was assessed with transwell assay. The number of cells that migrated through membrane into the lower chamber were significantly lower in the stalbly PCDH9-expressing cells than the vector control cells (P<0.001). On the contrary, in Hep3B cells, the number of cells that migrated through membrane into the lower chamber were significantly higher in both shRNAl and shRNA2(P<0.001). Similar results was showed in PLC/PRF/5cells. These results suggested that PCDH9indeed possessed the capability to inhibit HCC cells migration.5.PCDH9inhibits HCC cells epithelial to mesenchymal transition (EMT)During cell culture, we observed a loss of mesenchymal morphologic feature and re-emerging of epithelial features in PCDH9stable-expressing SNU-449cells. Such morphological changes prompted us to test the expression of a series of key EMT markers in the cells stably expression ectopic PCDH9. The quantificational RT-PCR revealed the relative reduction of mesenchymal markers including N-cadherin (CDH2,17.2fold), Vimentin (VIM,2.6fold) and Fibronectin (FN,281.5fold), as well as the increased expression of the epithelial marker such as E-cadherin (CDH1,13.5fold), Occludin (OCLN,6.5fold), Desmoplakin (DSP,3.7fold) and Keratin18(KRT18,2.2fold) in stably expression ectopic PCDH9compared with vector control SNU-449cells (P<0.001). Notably, the expression of the EMT-associated transcription regulators Snail1(Snail), which repressed E-cadherin transcription, was also decreased (6.5fold) in ectopic PCDH9expressing cells (P<0.001). The up-regulation of E-cadherin and the down-regulation of Snail and N-cadherin in those PCDH9re-expressed cells were further confirmed by western blot assay.Snail1plays a key role in the complex network orchestrated the epithelial-mesenchymal transitions, the rapid turnover of Snail1protein is GSK-3P mediated phosphorylation dependent. To determine whether GSK-3β involved in the Snail1down regulation during HCC cell mesenchymal-epithelial transitions induced by PCDH9re-expression, Western blot assay was conducted to detect the phospho-GSK-3p (Ser9), the inactive status of GSK-3β. Its showed that the level of phospho-GSK-3β (Ser9) was decreased in SNU-449cells with ectopic PCDH9expression, suggested that PCDH9could active GSK-3β signaling pathway. Since both Akt and Erk can phosphorylate GSK-3β at Ser9residue, we therefore detected the activation of Erk and/or Akt kinase in those PCDH9re-expressing cells. The down-regulation of phospho-Akt (Ser473) and phospho-Erks (Thr202/Tyr204) revealed by western blot strongly suggested involvement of Akt and ERKs kinases in PCDH9-mediated activation of GSK-3β.To verify activation of GSK-3β is required for PCDH9-inhibited EMT, we treated cells with GSK-3β inhibitor lithium chloride (LiCl)24h before harvest. Pre-treatment of SNU-449cells with LiC1could reverse the PCDH9mediated Snail down-regulation, accompanied with the increase of N-cadherin and decrease of E-cadherin levels. Meanwhile, we found that pre-treatment of SNU-449cells with LiCl could eliminate PCDH9overexpressing induced EMT markers change, the Iog2ratio of EMT markers was not more than1or-1.Moreover, to further testify that PCDH9inhibit Snail through activation of GSK-3β, we detected the expression of EMT markers and GSK-3β signal pathway after knockdown endogenous PCDH9in Hep3B and PLC/PRF/5cells. As shown in Figure4B, both in Hep3B and PLC/PRF/5cells, shRNA1and shRNA2could enhance the expression of Snail and N-cadherin, down-regulated E-cadherin. Moreover, knockdown endogenous PCDH9in Hep3B and PLC/PRF/5cells could up-regulation of phospho-Akt (Ser473) and phospho-Erks (Thr202/Tyr204) by western blot and up-regulation the level of phospho-GSK-3β(Ser9), suggested that knockdown endogenous PCDH9could active Erk and Akt signaling pathway and inhibite GSK-3β signaling pathway.What’s more, the quantificational RT-PCR revealed that knockdown endogenous PCDH9(shRNAl and shRNA2) in Hep3B cells could reduction of epithelial marker such as E-cadherin (CDH1,6.5and4.7fold), Occludin (OCLN,4.5and2.3fold), Desmoplakin (DSP,3.6and2.8fold) and Keratin18(KRT18,3.1and2.6fold) as well as the increased expression of the mesenchymal markers including N-cadherin (CDH2,4.5and3.3fold), Vimentin (VIM,3.0and2.5fold) and Fibronectin (FN,17.6and11.2fold). Notably, the expression of Snail was also increased (8.6and5.4fold) in ectopic PCDH9expressing cells (Figure4D). Still, knockdown endogenous PCDH9(shRNAl and shRNA2) showed similar results in PLC/PRF/5cells. These results suggest that PCDH9may inhibit Snail through suppressing Erk/Akt and activating of GSK-3β.Conclusion:PCDH9acts as a novel tumor suppressor candidate gene in HCC. The genetic and/or epigenetic aberration mediated PCDH9down-regulation was of clinical significance in HCC. PCDH9may inhibit EMT through suppressing Erk and or Akt and activating of GSK-3β. |
PDF Full Text Request