【作者】 黄静；

【导师】 陈龙华；

【作者基本信息】 南方医科大学 ， 肿瘤学（专业学位）， 2013， 博士

【摘要】 研究背景和目的原发性肝细胞癌(Hepatocellular caicinoma,HCC；简称肝癌)是世界上最常见恶性肿瘤之一。其死亡率位居全球恶性肿瘤死亡的第3位,在我国已上升至第2位,总体发病率和死亡率呈上升趋势。尽管近半个世纪来早期诊断、手术切除及药物治疗等方面不断发展,但肝癌患者的5年生存率仍然很低。临床上有54.1%-61.5%的HCC在行根治性切除后5年复发。复发和转移仍然是影响肝癌预后的主要原因。因此,探讨与肝癌发生发展相关因素,寻找预防和治疗的有效途径,提高病人的生存率,降低死亡率是目前肝癌研究的重点。肿瘤发生发展是一个多步骤、多阶段的复杂过程,包括肿瘤细胞从原发灶脱落,侵入血管或淋巴管,迁移、粘附于适宜部位,诱导肿瘤血管形成,对抗宿主抗肿瘤免疫,最终在远处形成转移灶。在这一过程中,涉及到多基因的参与。目前研究肿瘤发生发展机制大多关注癌基因、抑癌基因、小分子RNA等单个或多个分子的作用和功能,而蛋白质作为基因编码产物,是生命活动的最终执行者,其在肿瘤中的表达、对肿瘤发生发展所起的作用,仍有待阐明。近年来开展的蛋白质组学技术为揭示肿瘤发生发展机制带来了新的希望,而对蛋白质的生物学功能研究也成为当前的研究热点。Tiaml (T lymphoma invasion and metastasis1, T淋巴瘤侵袭转移基因)蛋白是我们在前期研究中发现的肝癌转移相关蛋白。在前期研究中,我们通过对213例具有完整临床随访资料的肝癌组织标本进行了免疫组织化学检测,结果显示：Tiaml在肝癌组织中高表达,且表达水平与肝癌患者预后、转移密切相关,可以作为判断肝癌患者预后的指标。Tiaml在8种肝癌细胞中的表达水平明显高于人正常肝细胞株。我们以人原发性肝癌高、低转移细胞株(HCCLM6, MHCC97L)为研究对象,采用基因转染和RNAi技术从细胞和动物水平证明了Tiaml过表达能促进肝癌增殖、侵袭、转移,相反沉默内源性Tiaml表达降低了肝癌增殖、侵袭、转移能力。研究结果为阐明肝癌转移机制及寻找预测肝癌转移的潜在标志物提供了理论依据。寻找Tiaml蛋白的相互作用蛋白是进一步理解Tiaml蛋白功能的一个重要途径。本研究拟在前期研究的基础上,利用酵母双杂交技术筛选Tiaml相互作用蛋白,为阐明肝癌发生、发展机理,寻找肿瘤早期诊断和治疗的靶标提供重要理论依据。方法1.酵母双杂交筛选Tiaml相互作用蛋白1.1构建酵母双杂交诱饵表达载体首先应用基因重组技术,以商品化的人Tiaml cDNA克隆为模板,构建Tiaml蛋白酵母表达质粒pGBKT7-Tiam1/C1199。将其转化到酵母菌株AH109中,检测Tiaml蛋白在酵母菌株中的表达情况,分析有无毒性和自激活活性。1.2利用cDNA文库筛选Tiaml相互作用蛋白利用经典的PEG/LiAc酵母转化法,采用大规模转化技术,以pGBKT7-Tiam1/C1199重组蛋白为诱饵,筛选"Universal Human Mate&Plate"文库。选取初筛获得的阳性克隆,进行质粒抽提、分离文库质粒和诱饵质粒、回复杂交验证。将获得的真阳性克隆进行测序分析和在NCBI数据库中进行Blast比对。2.交联免疫沉淀技术验证候选蛋白与Tiaml之间存在相互作用利用基因重组技术,分别构建了带有Flag和HA表位标签的候选蛋白和Tiam1-C1199真核表达载体。两种载体共转染人胚肾细胞HEK293T,加入含1%甲醛的PBS交联10min,超声裂解获取细胞总蛋白,将蛋白、抗体与Protein G共同孵育,通过anti-HA和anti-Flag的抗体分别进行免疫复合物沉淀,再以anti-Flag和anti-HA抗体进行western blot检测,确定Tiaml和SETDB1、ZNF307是否存在相互作用。3.肝癌组织中SETDB1的表达检测及临床意义利用免疫组织化学的方法检测87例具有5年完整临床随访资料的肝癌组织标本中SETDB1的表达情况,探讨SETDB1与肝癌发生发展的关系。4. SETDB1表达沉默对肝癌细胞生物学特性的影响应用荧光定量PCR和Western blot技术分别从mRNA和蛋白水平检测SETDB1在肝癌细胞系中的表达。使用Designer3.0(Genepharma)软件进行设计,合成4条SETDB1基因的特异性干扰序列A、B、C、D。将寡核苷酸链退火成双链DNA后,与线性化的慢病毒载体相连,构建SETDB1慢病毒干扰载体,将慢病毒载体质粒与辅助质粒共转染人胚肾HEK293细胞包装慢病毒,收集病毒上清,进行病毒滴度测定。用含SETDB1特异性干扰片段的慢病毒感染内源性高表达SETDB1的肝癌细胞株,流式细胞仪进行荧光分选,荧光定量PCR、western blot检测转染各干扰片段细胞中SETDB1基因的干扰效率,挑选干扰效率最高的细胞株作为体外功能试验的研究对象。利用CCK8法、平板克隆形成实验及流式细胞周期技术检测SETDB1沉默后对细胞体外生长和增殖能力的影响；利用划痕实验及Transwell小室分析体外迁移和侵袭能力的变化。5.统计学分析采用SPSS13.0对数据进行统计学分析。采用χ2检验对SETDB1蛋白表达与各临床病理参数之间的关系进行统计学分析；对SETDB1蛋白表达及其它临床病理参数与生存期的关系进行统计学分析采用Kaplan-Meier法和Cox风险比例回归模型。分别采用析因分析、重复测量资料的方差分析方法检测转染前后体外增殖能力和迁移能力的变化；采用单因素方差分析荧光定量PCR、平板克隆形成实验、细胞周期实验、侵袭实验结果,检验各组细胞之间的统计量是否存在差异,方差齐时用LSD法,方差不齐时用Dunnett’s T3法。均以P<0.05为差异具有统计学意义。结果1. Tiaml相互作用蛋白的筛选1.1成功构建诱饵蛋白pGBKT7-Tiam1/C1199将pGBKT7-Tiaml/C1199诱饵蛋白质粒转化酵母菌株AH109,应用Western Blot技术,检测Tiaml蛋白在酵母菌株中的表达良好,诱饵蛋白在营养缺陷平板上可以生长,且无毒性和自激活活性,可用于下一步的文库筛选。1.2酵母双杂交筛选Tiaml相互作用蛋白以pGBKT7-Tiaml/C1199为诱饵,筛选"Universal Human Mate&Plate"文库,共获得24个与Tiaml诱饵蛋白相互作用的真阳性克隆,经测序证实及Blast比对,它们分别是OSBPL1A、ZKSCAN4(又名ZNF307)、FNDC3B、SRSF5、 SYCP1、SETDB1。经文献挖掘和生物信息学分析,选取SETDB1和ZNF307进行下一步的验证实验研究。2. Tiaml相互作用蛋白的鉴定通过交联免疫沉淀实验,在体内分别验证Tiaml与SETDB1及ZNF307的相互作用。结果表明,经anti-Flag抗体沉淀的蛋白复合物中检测到Tiaml和Flag-SETDB1蛋白,说明交联免疫共沉淀成功；相反用anti-HA抗体做免疫沉淀,在沉淀的蛋白复合物中未检测到SETDB1蛋白的表达,而HA-Tiaml表达可以检测到。在ZNF307与Tiaml的免疫共沉淀实验中,我们未获得阳性结果。这些结果证明了SETDB1和Tiaml在细胞体内确实存在相互作用。3.肝癌组织中SETDB1的表达检测及临床意义免疫组化结果显示,SETDB1蛋白表达的阳性信号主要定位在细胞浆内,仅在个别病例中发现间质中有表达。SETDB1蛋白在癌旁组织中呈高表达(阳性～强阳性,49/54例),而在癌组织中呈低表达趋势(阴性～弱阳性)。SETDB1蛋白表达水平高低与患者的性别、年龄、肿瘤大小、HBsAg.血清AFP、肿瘤转移及复发无明显相关性(P值分别为0.140,0.090,0.927,0.245,0.091,0.102,0.245,0.068),而与肿瘤分化程度及肝硬化相关(P值分别为0.008,0.006)。生存分析结果显示：SETDB1蛋白高表达患者中位生存时间为48.0个月,而SETDB1蛋白低表达患者的中位生存时间为19.0个月,可见SETDB1低表达与患者的不良预后相关(Log-rank检验：x2=4.106,P=0.043)。将临床各病理参数导入Cox比例风险回归模型进行全变量模型检验,结果提示：年龄、HBsAg、血清AFP、转移、复发是影响患者生存时间的主要因素(P值分别为0.009,0.010,0.018,<0.001,0.010)。然而就目前COX回归分析结果而言,SETDB1表达水平尚不能作为预测患者不良预后的独立指标。4.荧光定量PCR、Western blot方法检测SETDB1在肝癌细胞株中的表达SETDB1蛋白在8种原发性肝癌细胞株中均有表达,表达水平各异,其中在MHCC97L和Bel-7402中表达量较低,在SMMC-7721和HCCLM3中表达水平较高。蛋白表达趋势跟mRNA水平的表达趋势基本一致,最终选取内源性高表达的HCCLM3肝癌细胞株作为下一步实验的研究对象。5. SETDB1表达沉默对肝癌细胞生物学特性的影响利用CCK8检测SETDB1表达沉默对肝癌细胞株体外生长能力的影响,结果表明：随着时间的变化,三组细胞各时间点生长速度变化趋势不同。与野生型M3细胞和空白对照组M3/mock细胞相比,特异性沉默内源性SETDB1表达后的M3/SET-细胞生长减慢,不同细胞组间差异有统计学意义(F=424.808,P<0.001),不同测量时间之间差异有统计学意义(F=1100.419,P<0.001),二者有交互效应(F=44.516,P<0.001)。平板克隆形成实验的结果显示M3/SET-组细胞的增殖能力弱于其他两组细胞,差异有统计学意义(F=198.358,P<0.001).流式细胞仪分析细胞周期变化,结果显示,SETDB1基因表达沉默后,细胞发生了G1期细胞比例明显增加,G2期细胞比例相应减少,S期细胞比例变化不明显。提示SETDB1基因表达沉默通过细胞G1期阻滞,减少合成期的细胞比例,从而抑制了细胞增殖,这与CCK8和平板克隆形成实验的结果相呼应。利用划痕实验检测SETDB1表达沉默对细胞迁移运动的影响,结果显示：划痕72h后,SETDB1稳定干扰组的细胞迁移能力低于其他两个对照组细胞,差异具有统计学意义(交互效应F=48.784,P<0.001)。说明SETDB1表达沉默能降低肝癌细胞的迁移能力。利用Transwell小室进行体外侵袭实验,在相同条件下比较SETDB1稳定干扰和对照组细胞穿透基底膜的能力。细胞侵袭42小时后观察三组细胞均可在聚碳酸酯膜上见到发生形态变性的细胞,分析显示,M3/SET-组穿过聚碳酸酯膜的细胞数少于其他两组,差异有统计学意义(F=721.0,P<0.001),说明SETDB1表达沉默抑制了肝癌细胞的体外侵袭能力。结论1.利用基因重组技术,成功构建Tiaml-C1199诱饵表达质粒pGBKT7-Tiam1/C1199。利用经典的PEG/LiAc酵母转化法,采用大规模转化技术,以pGBKT7-Tiam1/C1199为诱饵,筛选"Universal Human Mate&Plate"文库,经回复杂交验证及测序比对分析,共获得6个Tiaml相互作用蛋白。2.成功构建带不同标签的蛋白表达载体,共转染人胚肾细胞HEK293T,通过交联免疫沉淀技术鉴定,证明SETDB1与Tiaml在哺乳动物细胞体内确实存在相互作用。3.免疫组织化学检测和临床资料分析结果表明：SETDB1蛋白表达水平与肿瘤分化程度及肝硬化相关。SETDB1低表达与患者的不良预后相关。SETDB1表达水平、肿瘤分化程度、复发、转移是影响肝癌患者生存时间的因素。然而COX回归分析提示SETDB1表达水平尚不能作为预测患者不良预后的独立指标。4. SETDB1基因表达沉默可以抑制肝癌细胞的体外增殖、迁移、侵袭能力。本研究的创新之处1.利用酵母双杂交系统结合基于蛋白质水平的交联免疫共沉淀技术分析筛选鉴定出肝癌转移相关蛋白Tiaml的相互作用蛋白SETDB1,为研究肝癌转移的分子作用机制提供了新思路。2.利用慢病毒载体和RNAi技术建立了稳定的SETDB1基因表达沉默的肝癌细胞株,初步证实了SETDB1基因与肝癌发生发展的关系,为进一步研究Tiam1-SETDB1在肝癌发生发展中的参与的调控通路及其作用机制提供了实验依据。本研究的不足之处1.现有的验证实验尚不能说明SETDB1与Tiaml之间的相互作用是间接还是直接的,后续研究增加了体外GST pull down和免疫荧光共定位实验验证。2. SETDB1表达水平能否作为临床肝癌患者预后的独立预测指标,仍有待在大样本的群体中证实。3. SETDB1基因在肝癌发生发展中的生物学功能尚需体内试验进一步证明。更多还原

【Abstract】 Background and ObjectiveHepatocellular carcinoma (HCC), one of the most common cancers in the world, is the third leading cause of cancer-related deaths worldwide and the second in our country. Population-based studies show an increasing incidence and mortality. Although early diagnosis, surgery and drug therapy have been improved steadily, the five-year survival rate remains low. There are approximately54.1%-61.5%of HCC patient relapse with radical excision. Metastasis and recurrence are the two major factors affecting the outcome. It is an urgent task to work out the metastasis-associated factors and find out the preventive and therapeutic methods so that to decrease mortality and improve overall survial.The development and progression of malignant is a complex multistep process, involving the tumor cells shed from the primary tumor, invade into the blood vessels or lymphatic vessels, migrate and adhere to appropriate site, induce tumor angiogenesis, resistant to antitumor immunity and ultimately form metastases in the distance. There are multiple genes involved in this complex process. At present, most of the researchs associated with the tumor pathogenesis focuse on the function of some single or multiple molecules such as oncogenes, anti-oncogene, small RNA. As gene encoding product, protein acts as the final execution of life activities and plays important roles in tumorigenesis. It remains to elucidate the expression and role of protein in tumor pathogenesis. The proteomics technology carried out in recent years has brought new hope for revealing the mechanisms of tumor development, and research on the biological function of the protein has also become a current research focus.Tiaml protein is a novel metastasis-related protein for HCC found in our previous study. We previously examined the expression of Tiaml in213primary HCC tissue samples with5-year complete clinical follow-up data by IHC. As the result shows, the Tiam1protein expression was strong in HCC tissues. Tiaml is considered to be correlated to the metastasis and overall survial of HCC patients, thus could be served as a biomarker for predicting outcome for HCC patients. Tiaml expression in HCC cell lines was significantly higher than the normal human liver cell line. Focused on the high and low metastastic HCC cell lines (HCCLM3, MHCC97L), we determined that Tiaml overexpression could promote cell proliferation, invasion and metastasis by using gene transfection and RNAi technology. On the contrary, Tiaml knockdown results the reverse. Our study should clarify the molecular mechanism for HCC metastasis and provides a theoretical basis for potential markers that for predicting HCC metastasis.Searching for Tiaml interacting protein is an important way to understand Tiaml protein function. Based on our previous study, we attend to screen Tiaml interacting protein by using yeast two-hybrid, thus to clarify the potential mechanism for HCC development and provide important theoretical basis for early diagnosis and therapeutic targets.Methods1. Screening of protein interacting with Tiaml by yeast two-hybrid1.1Construction of a yeast two-hybrid expression vector with bait protein A recombinant plasmid pGBKT7-Tiaml/C1199for Tiaml expressing in yeast cells was constructed by recombinant gene technique. C1199code domain sequence of Tiaml was amplified from a commercial Tiaml cDNA clone which was then cloned into pGBKT7vector and confirmed by sequencing. Yeast strain AH109cells were transiently transformed with pGBKT7-Tiaml/C1199plasmid. Western blot was performed to confirm that if Tiaml protein could be normally expressed in Saccharomyces cerevisiae without toxicity and autoactivation(Strain AH109).1.2Screening of protein interacting with Tiaml in cDNA library.We used pGBKT7-Tiaml/C1199as a bait to screen the "Universal Human Mate&Plate" Library by performing the classic PEG/LiAc yeast transformation method, and large-scale conversion technologies. Partial positive clones were selected for plasmid extraction, separation library plasmid (AD-X) and bait plasmids, rotary hybrid verification. Some true positive clones were obtained. These positive clones were sequenced and analyzed in the NCBI database.2. Crosslinking immunoprecipitation was applied to validate the interaction between candidate proteins and TiamlBy recombinant DNA technology, we constructed some different epitopelabel (Flag and HA) candidate protein and Tiaml recombinant expression vectors. The recombinant plasmid with different epitopelabel and Tiaml-C1199recombinant plasmids were co-transfected into human embryonic kidney cells HEK293T. Cells were fixed by10ml of1%formaldehyde in PBS at room temperature for10min. sonicated for total cellular protein. Protein, antibody and Protein G were incubated together. Anti-HA and anti-Flag antibodies were used respectively for the recipitation of immune-complexes. And then the immune-complexes were detected by anti-Flag and anti-HA antibody in western blot assay, so that to confirm whether Tiaml is really interect with SETDB1and ZNF307protein. 3. Detection of SETDB1expression in HCC tissues and its clinical significanceSETDB1expression were detected by IHC in87HCC patients’ tissue samples which with complete clinical follow-up data, and was analysed to explore its role in HCC development.4. Lentivirus-mediated silencing of SETDB1and the influence on human HCC cell line.Protein and transcriptional expression of SETDB1were detected in several HCC cell lines by real time PCR and western blot. Four short hairpin RNAs that specific targeted against SETDB1were designed by Designer3.0software (Genepharma).The oligonucleotides were annealed into double-stranded DNA, and then cloned into linearized lentiviral vector to construct an SETDB1interference lentiviral vector. The lentiviral vector plasmid and the helper plasmids were co-transfected into HEK293cells. The viral supernatant was collected and titered.HCC cell lines which is endogenous SETDB1high expressed were transfected with lentiviral containing SETDB1-shRNA. Flow cytometry were used for cell selecting with EGFP gene reported. Real time PCR and western blot were performed to test the RNA interference efficiency of SETDB1. The transfected cell line which exhibited the higest RNA interference efficiency was selected for further study.We employed CCK8assay, plate colony formation and flow cytometry to detect the effect of SETDB1on cell growth and proliferation. As well as we used wound-healing assay and Transwell inserts to compare the migration and invasion ability between the treat and control groups.5. Statistical analysisDatas are analyzed by SPSS13.0software packet. The χ2test (chi-square test) was used to analyzed the relationship between SETDB1expression level and clinicopathological parameters, while Kaplan-Meier survival analysis and Cox proportional-hazard model were applied for analysis of the correlationship between SETDB1expression/other clinicopathological parameters and survival time.The in vitroproliferation and migration ability of different treated cells were compared by factorial analysis and repeated-measures analysis respectively.The results of quantitative PCR, colony formation assay, cell cycle assay, invasion assay were analyzed by One-way ANOVA analysis, with LSD method and Dunnett’s T3method were used according to the homogeneity of variance test. P<0.05are considered to be statistically significant different.Results1. Tiaml-interacting protein by yeast two-hybrid screening1.1A bait protein, pGBKT7-Tiaml/C1199, was successfully constructed.The Tiaml/C1199truncated protein was expressed well without either toxicity nor auto-activating ability, when pGBKT7-Tiam1/C1199plasmid was transduced into AH109yeast strain. So that this bait could be used for the yeast two-hybrid screening.1.2Tiaml-interacting protein was screened by yeast two-hybridWe used pGBKT7-Tiaml/C1199as a bait to screen the "Universal Human Mate&Plate" Library. And finally we got a total of24positive clones, those were were sequenced and analyzed in the NCBI database, exactly match to6known proteins, including OSBPL1A, ZKSCAN4,(also known as ZNF307), FNDC3B, SRSF5, SYCP1and SETDB1. After literature mining and bio informatics analysis, we select SETDB1and ZNF307for further verification study.2. Identification of Tiaml-interacting proteinCrosslinking immunoprecipitation was applied to validate the interaction in v/vobetween candidate proteins (SETDB1and ZNF307) and Tiaml.The results showed that Tiaml and Flag-SETDB1protein could be detected by corresponding antibody when we do the IP by anti-Flag antibody. This result could prove a real interaction between Tiaml and SETDB1. However, we couldn’t get positive result when IP with anti-HA antibody in other direction. As far as now, we haven’t gotten a real positive result that could confirm the interaction of ZNF307and Tiaml.3. Expression of SETDB1in HCC and its clinical significanceImmunohistochemical results showed that SETDB1protein positive signals mainly localized in the cytoplasm, interstitial expression found only in isolated cases. We observed that SETDB1protein in tumor tissues showed lower expression level (positive to strongly positive,49in54) than their adjacent tissue (negative to weak positive). Results showed that SETDB1expression does not correlate with any of the clinicopathological parameters such as sex, age, tumor size, HBsAg, serum AFP, metastasis and recurrence (P value are0.140,0.090,0.927,0.245,0.091,0.102,0.245,0.068, respectively), whereas correlate with the tumor differentiation and cirrhosis of the liver.Our Kaplan-Meier survival analysis revealed that patients with SETDB1overexpression exist a longer median survival time (48m versus19m), which was significantly different (Log Rank=4.106, P=0.043<0.05). By univariate analysis of Cox proportional-hazard model, SETDB1low expression is closely related to the poor prognosis of the patients. Age, HBsAg, serum AFP, recurrence, and metastasis are associated with the survival of HCC (P value are0.009,0.010,0.018,<0.001,0.010, respectively). However, SETDB1expression could not server as an independent predictor for prognosis as the results showed in Cox proportional-hazard model.4. SETDB1expression in HCC cell lines were detected by real time PCR and western blotWe found that SETDB1expressed differently in8primary liver cancer cell lines. SETDB1mRNA expression in MHCC97L and Bel-7402are lower than that in SMMC-7721and HCCLM3. The result of western blot is quite the same with real time PCR. Finally, we selected HCCLM3, which is SETDB1endogenous high expressed for further study.5. Effects of SETDB1gene silencing on the biological features of HCC cells.We employed CCK8assay to detect the effects of SETDB1on cell growth. By comparing thegrowth curves of M3/SET-and its controls (M3parental and M3/m ock). As time passed by, M3/SET-was found to significantly suppress cell growth, there are significantly different among the three groups (different groups:F=424.808, P<0.001; different time piont:F=1100.419, P<0.001, cross action:F=44.516, P<0.001).Similarly,SETDB1silencing significantly reduces colony formation ability of HCCLM3cells relative to parental cells and cells transfected with empty vector (F=198.358,P<0.001).Flow cytometry revealed that endogenous SETDB1silencing could arrest Gl phase of cell cycle, which directly result in a reduction in proliferation. In a wound-healingassay, evident suppression in the wound closure rate wasobserved in suppression cells at72hr compared with the controls (cross action:F=44.516, P<0.001). The number of cells that passed through Matrigel after incubated for42h was as nearly six-fold lower than those in the mock and parental cell groups (F=721.0, P <0.001). All of these reveal that SETDB1silencing could reduce cells’ proliferation, migrate and invasive ability in vitro.Conclusions1. We successfully construct bait plasmid pGBKT7-Tiaml/C1199, by using traditional PEG/LiAc method and large tlassic PEG/LiAc yeast transformation method, and large-scale conversion technologies, we screened the "Universal Human Mate&Plate" library, and confirmed by reverse hybrid assay, and we finally obtained6positive potein interacting with Tiaml. 2. With plasmids co-transfected HEK293T cells, we indentified the interaction between Tiaml and SETDB1intracellular through Crosslinking IP.3. IHC assay and statictic analysis were performed to explore the relationship between SETDB1and HCC. By univariate analysis of Cox proportional-hazard model, SETDB1low expression is closely related to the poor prognosis of the patients. Age, HBsAg, serum AFP, recurrence, and metastasis are associated with the survival of HCC patients. Multivariate analyses revealed that Age, HBsAg, recurrence, and metastasis but SETDB1expression may play a role in predicting the overall survival in HCC patients. However, SETDB1expression could not server as an independent predictor for prognosis in this study.4. SETDB1silencing could suppress HCC cells proliferation, migration and invasion ability.Innovations1. We screened and identified SETDB1interacte with metastasis-related protein Tiaml by using the yeast two-hybrid system combined with the Cross-linked immune precipitation technology. It provides new opportunities for researching the molecular mechanisms of metastasis of HCC.2. We successfully constructed a cell line stably express low SETDB1, and confirmed the function of SETDB1in HCC, which should provide an new insight for clarifying the molecular mechanism of Tiaml-SETDB1in the pathogenesis of HCC.Weakpoints1. All of our work are not sufficient for elucidating the direct or indirect interaction between Tiaml and SETDB1. GST pull down and Immunofluorescent techniques could compensate it.2. This study is still unable to confirm the value that expression level of SETDB1serves as a prognosis independent predictor of HCC, which need a further validation in a larger set of sample.3. The biological function of SETDB1gene in liver cancer is required in vivo further experimental verification.更多还原