Inhibitory Effects Of Silencing Osteopontin Expression By Lentiviral-mediated RNAi On The Tumor Growth And Metastasis Of Human Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) is one of the most common and aggressivemalignancies worldwide, ranking the fifth most important cancer in terms of numbersof cases and the third in terms of cancer mortality. And it has become the secondcancer killer in China since 1990s. Unfortunately, despite efforts to improve itsprognosis, the overall survival of patients with HCC is still dismal. Metastaticrecurrence after HCC resection is one of the major obstacles to prolonging survival.Therefore, there is a tremendous interest and urgency to search for molecules relatedto HCC metastasis that would provide new predictors for HCC metastasis as well asnew targets for intervention.Osteopontin (OPN), a secreted phosphoglycoprotein, is a promoting factor fortumorigenicity and metastasis of cancer cells in several types of tumors through avariety of mechanisms. OPN binds with several integrins and CD44 variants in anRGD sequence-dependent and-independent manner. In our previous work, wecollaborated with the National Cancer Insititute (NCI) of USA to compare thedifference in gene expression profiles between HCCs without and with accompanyingintra-hepatic metastases using a cDNA microarray of. 9180 cDNA clones. We foundthat OPN was one of the leading genes that over-expressed in the HCCs withmetastasis. An OPN-neutralizing antibody efficiently blocked in vitro invasion and invivo pulmonary metastasis of HCC cells. Moreover, we also found that preoperativeplasma OPN level can be used as a predictive marker for tumor recurrence andprognosis of patients with HCC after operation. These suggest that OPN is associatedwith metastasis of HCC. However, the mechanism by which OPN mediatesprogression and metastasis of HCC remains unknown. So, in this study, we tried tofind out what happened with the metastatic HCC cell line (HCCLM3) if its highendogenous OPN level was knocked down using lentiviral-mediated RNAinterference, and explore the possible mechanisms. PARTâ… Construction and selection of Lentiviral-mediated RNAiexpression vectorsThe pcDNA ^TM6.2-GW/EmGFP-miR expression vectors contain adouble-stranded oligonucleotide (ds oligo) encoding a pre-miRNA sequence forexpression in mammalian cells using a RNA Polymeraseâ…¡(Polâ…¡) promoter, thehuman cytomegalovirus (CMV) immediate early promoter. The pDONR^TM221 vectoris used as an intermediate to transfer the pre-miRNA expression cassette into thelentiviral expression plasmid (pLenti6/V5-DEST) using Gateway Technology. Weperformed the Rapid BP/LR recombination reaction between pDONR^TM221,pcDNA^TM6.2-GW/EmGFP-miR and pLenti6/V5-DEST to generate the pLenti6/V5-GW/EmGFP-miR expression vectors. The sequence for construction of four OPN(GenBank accession number NM₀00582) pre-miRNA and one control pre-miRNAwere as follows:Packaging vector (9μg) (pLP1, plp2, pLP/VSVG), 3μg pLenti6/V5-GW/EmGFP-miR expression plasmid with OPNi pre-miRNA or control pre-miRNAwere cotransfected into 293FT packaging cells with Lipofectamine 2000 (36μl). After48 h the virus-containing culture supernatants were centrifuged to pellet the cell debris and stored in 1 ml aliquots at -80℃. Before proceeding to transduce theHCCLM3 cell line and express the miRNA for RNAi analysis, we determined the titerof lentiviral stock using EmGFP tittering method, calculated the EmGFP lentivirustiters from the dilutions at which the percentage of EmGFP-positive cells fall withinthe range of 1-30ï¼…. This was to avoid analyzing dilution samples containing multipleintegrated lentiviral genomes, which may result in an underestimate of the viral titer.Titer is expressed as transducing units (TU)/ml. The following formula was used tocalculate the titer: [F×C/V]×D(F=frequency of GFP-positive cells (percentageobtained divided by 100) C=total number of cells in the well at the time oftransduction V=volume of inoculum in ml D=lentivirus dilution). The lentiviraltiter for this study was 1×10⁸ TU/ml. To obtain optimal expression of miRNA andtherefore, the highest degree of target gene knockdown, the lentiviral constructionwas transfected into mammalian cell line using a suitable MOI. MOI was defined asthe number of virus particles per cell and generally correlates with the number ofintegration events and as a result, expression. Typically, miRNA expression levelsincreased as increasing MOI. The MOI was determined by detecting the EmGFPfluorescent protein, and the suitable MOI was 30 for HCCLM3 cell line. HCCLM3cells were transfected with lentiviral vectors in the presence of polybrene (6μg/ml).After day 3, the medium was replaced with fresh, complete medium containg0.5mg/ml Blasticidin to select for stably transfected cells. Four stably transfected cells(Lenti.OPNi-1, Lenti.OPNi-2, Lenti.OPNi-4, Lenti.OPNi-3M) were obtained.Interestingly, only Lenti.OPNi-3 could inhibit the proliferation of HCCLM3significantly, and no Blasticidin-resistant colonies were identified. Real-time PCRassays showed that Lenti.OPNi-1, Lenti.OPNi-2, Lenti.OPNi-3, Lenti.OPNi-4 andLenti.OPNi-3M could suppress the OPN expression by 60ï¼…, 72ï¼…, 90ï¼…, 52ï¼…and14ï¼…, respectively. In addition, no significant induction of P-elF2αwas found, whichindicated that the inhibitory effect of RNAi was not through the activation ofIFN-response genes.PARTâ…¡Inhibitory effect of silencing OPN expression on the in vitroproliferation and invasive abilities of human HCC cells 1. Effect of OPN silencing and HCC proliferationOf these miRNA lentiviral vectors, we selected Lenti.OPNi-2 and Lenti.OPNi-3which mediated efficient OPN suppression to evaluate effects of OPN knockdown onin vitro proliferation of HCCLM3 cells. Cell proliferation was measured by cellularuptake of MTT (3,-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide).Each assay was performed three times. Lenti.OPNi-3, but not Lenti.OPNi-2, couldinhibit the cell proliferatin significantly (p＜0.001). Lenti.OPNi-3 induced a significantdecrease in MEK and ERK1/2 phosphorylation (p＜0.001), while in cells transfectedwith Lenti.OPNi-3M there were no alteration in the level of OPN and in the activationof downstream pathways (p=0.482). In addition, Lenti.OPNi-3 did not induceaspecific downregulation of endogenous genes such as GAPDH and NF-κB. Incontrast, the proliferation rate of HepG2 without OPN overexpression was lesssensitive to the inhibitory effect of the OPN knockdown (p=0.216), and no changewas observed in the phosphorylation of MEK and ERK1/2 (p=0.127). Theconstitutively low-level expression of the OPN in the HepG2 cells may explain thereason for the HepG2 resistance to the OPN RNAi. Similar result was observed inCCL13 cells. Thus, lentiviral mediated OPN knockdown does not appear to affectcells with low OPN expression. To further investigate whether OPN inducesphosphorylation of MEK and then phosphorylation of ERK1/2 in HCCLM3 cells, thecells transfected with Lenti.OPNi-3 were treated with 10μM exogenous recombinantOPN, pretreated with specific MEK inhibitor U0126 (20μM), or treated with OPNalone. Cell lysates were analyzed by Western blot at different time points. The dataindicated that MEK and ERK were readily activated in OPN-stimulated cells, peakingat 20 minutes, but not in the cells treated with U0126.2. Effect of OPN silencing and tumor invasionWe evaluated whether the suppression of OPN expression would alter the in vitrometastatic phenotype of HCCLM3 cells. Cell invasion was measured using Matrigelcoated transwell-invasion chambers. Migrating numbers of HCC-LM3 cells wassignificantly inhibited by Lenti.OPNi-2 (9.2±1.3) and Lenti.OPNi-3 (7.8±1.5),compared with Mock (28.84±2.6) (p＜0.001). No significant difference in invasivenesswas observed between Mock and Lenti.OPNi-3M (p=0.221). Parallel experimentswere performed in triplicate and each filter was photographed at 200×magnification.Next, we evaluated the protein expression of several metastatic markers includingMMP-2, MMP-9 and uPA, which have been previously shown to be associated withOPN in models of breast cancer and melanoma. Interestingly, MMP-2 and uPA expression but not MMP-9 were decreased significantly in Lenti.OPNi-2 andLenti.OPNi-3(p＜0.001). Previous studies have shown that OPN stimulates activationof pro-MMP-2 through NF-κB in murine melanoma cells. To further determinewhether OPN have the same effect on translocation of p65 into the nucleus inHCCLM3 cell, the cells were treated with 10μM OPN in basal medium for 0-2h at37℃. Immunofluorescence analysis of NF-κB was performed at different time points.The data showed that OPN induces translocation of p65 into nucleus in atime-dependent manner. In the cells pretreated with OPN, the majority of p65 stainingresided in the cytoplasm upto 30 min. At 45 min and 70 min some nucleartranslocation of p65 were observed, peaking at 2h. Western blotting analysis gave afurther confirmation of translocation of p65 into the nucleus. In the OPN-untreatedcells, the p65 was localized mostly in the cytoplasm compared with the nucleus,whereas in the OPN-treated cells, it was translocated into the nucleus. In order tocheck whether OPN induced pro-MMP-2 production and activation in these cells, theconditioned medium was collected, and MMP-2 activity was analyzed by gelatinzymography. The levels of both the pro and active forms of MMP-2 in theOPN-treated cells were significantly higher compared with the levels of MMP-2 inOPN-untreated cells(p＜0.01).Inhibitory effect of silencing OPN expression on the in vivotumor growth and lung metastasis of human HCC cellsThe in vitro study supported that the hypothesis that OPN contributed totumorigenicity and metastatic capacity of HCCLM3 cells. To directly confirm this invivo, HCCLM3 cells (5×10⁶) were transfected with or without the lentiviral vectors,and then subcutaneously inoculated into nude mice. Tumor growth was monitored. At2 week after the injection, the tumors originating from the cells transfected withLenti.OPNi-2 and PBS were measurable. At 6 weeks, the animals were killed forethical reasons given the tumor volume, and then the tumors were excised from mice.HCCLM3 cells mock-transfected (PBS) or transfected with Lenti.OPNi-2 developedlarge tumors 3500mm³ within 6 weeks. However, growth of HCCLM3 cellstransfected with Lenti.OPNi-3 was markedly suppressed (p＜0.001), suggesting thatknockdown of OPN leads to significant inhibition of tumor growth in vivo. In the sixth week, pulmonary metastatic lesions were detected in every mouse inPBS group, with mainly gradeâ… -â…¡and some gradeâ…¢-â…£tumor clusters. The mice ofPBS group had an average of 14+2.4 tumor clusters per lung. However, Only 50% miceinfected with the Lenti.OPNi-3-treated cells developed lung metastases with mostly gradeâ… tumorclusters with a combined average of 1.2±1.3 tumor clusters per lung. The effect wasstatistically significant(p＜0.01). The same significant effect was also observed in themice infected with the Lenti.OPNi-2-treated cells(p＜0.01). We also monitored HCCLM3cells growth and metastasis with bioluminescence imaging every week. Interestingly,EmGFP fluorescent protein can be detected in the mice with no measurable tumormass, which suggests the proliferation of HCCLM3 cells transfected withLenti.OPNi-3 was inhibited significantly.Conclusions1. We constructed two lentiviral vectors (Lenti.OPNi-2 and Lenti.OPNi-3) whichcould significantly suppress the expression of OPN, and selected two stablytransfected HCCLM3 cell lines (LM3- Lenti.OPNi-2 and LM3- Lenti.OPNi-3M).2. Osteopontin could promote HCCLM3 cell proliferation through activation ofMEK/MAPK signaling pathways, and little basal expression of OPN is requiredfor HCCLM3 proliferation. However, the cells with constitutively low-levelexpression of OPN (HepG2 and CCL13) were less sensitive to the inhibitoryeffect of the OPN knockdown.3. Osteopontin stimulates HCCLM3 cell invasion through activation of NF-κB,MMP-2 and uPA. Both Lenti.OPNi-2 and Lenti.OPNi-3 can inhibit pulmonarymetastases of HCCLM3 cells significantly which suggest more basal expressionof OPN is required for HCCLM3 invasion than for HCCLM3 proliferation.Potential application of this study1. Long lasting RNAi-based gene knockdown can be achieved by lentiviral basedexpression systems driving the production of short hairpin RNA species (microRNA).Lentiviral-mediated gene silencing is a good way to assess one gene function.2. Lenti.OPNi-3 significantly inhibits proliferation and invasion of HCCLM3 withoverexpression of OPN, but does not appear to affect CCL13 with low OPN expression.Maybe OPN will become a new therapeutic target for highly metastatic HCC, and Lentiviral vectors provide some insight into vector tools of the bilolgical targettherapy in liver cancer.Novelties1. To the best of our knowledge, this is the first report showing that OPN expressionis critical for tumorigenesis and proliferatin of human hepatocellular carcinomacells since knocking down the expression abolishes tumorigenicity.2. In addition, our data for the first time show that different biological behaviors ofHCC have a different requirement for the basal expression level of OPN.