【作者】 周士新；

【导师】 陆祖宏；

【作者基本信息】 东南大学 ， 生物医学工程， 2005， 博士

【摘要】 恶性肿瘤细胞向其他组织或器官转移是肿瘤的一个基本特征,也是恶性肿瘤难以治愈的关键原因,人们目前对于肿瘤的转移机制和影响转移的基因了解的还不很清楚,特别是不同肿瘤细胞的转移机制和影响其转移的关键基因存在一定的差异。恶性肿瘤的转移涉及到多阶段改变的过程,其中包括:局部侵袭、血流转运、溢出、增殖、定着靶器官等,这些过程包含多种基因的复杂变化。我们利用从小鼠树突状细胞瘤(DCS)分离出来的肺转移能力明显差异的两株亚克隆细胞株,采用高密度的Affymetrix基因芯片检测这两株细胞的基因表达谱,用生物信息学的方法分析基因表达数据,获取表达显著差异的基因,进行了以下几方面的研究:1.小鼠树突状细胞瘤转移相关基因的研究小鼠树突状细胞瘤(DCS)是中国医学科学院(协和医科大学)基础医学研究所细胞中心所建的细胞系,他们用极限稀释方法从小鼠DCS母细胞系中分离到与同一母细胞来源的,转移能力明显差异的DD1和DG6细胞株,为消除细胞种群等其他因素的影响,专门研究DCS与肿瘤转移相关基因提供了良好的细胞模型。我们用高密度的Affymetrix基因表达芯片(小鼠MOE-430A),比较两株细胞基因表达的差异,高通量地筛选出可能与小鼠DCS肺转移相关的基因。结果发现,高转移细胞株DD1和低转移细胞株DG6在转录因子、RNA过程、细胞增殖和凋亡、分子转运和蛋白质修饰等方面均存在差异;通过RT-PCR和细胞免疫荧光试验,在RNA和蛋白质水平上我们验证了影响细胞分化的转录因子Ehf、凋亡相关基因Pedf、细胞粘附因子Nectin-3和趋化因子Mcp-1(CCL2)在这两株细胞中存在着显著的差异,提示这些基因可能与小鼠DCS肺转移相关。2. DNA甲基化与小鼠树突状细胞瘤的基因表达我们选择在高转移DD1和低转移DG6细胞株中表达差异较大一些转录因子Smcy,Ehf及与细胞凋亡Pedf、细胞粘附Nectin-3有关的基因,通过检测这些基因启动子区DNA甲基化情况,探讨是否由于DNA甲基化而控制这些基因的上调和下调,初步探寻影响小鼠DCS细胞基因表达的机制。我们采用亚硫酸氢盐处理高转移细胞株DD1和低转移细胞株DG6的基因组DNA后,用不含CG序列的特异性引物经PCR扩增出这几个基因的启动子序列,然后测序观察结果,CG序列中未甲基化的胞嘧啶C被转化胸腺嘧啶T,而甲基化的C保持不变。结果发现Pedf基因有一个位点的甲基化改变(C变为T);Smcy和Ehf基因启动子区的DNA甲基化模式没有明显改变,我们倾向认为,可能不是由于DNA甲基化控制这几个基因的高表达、低表达,而是其他原因如转录因子之间的相互作用影响这几个基因的上调或下调。3.小鼠树突状细胞瘤基因表达谱分析我们用Affymetrix基因表达芯片(小鼠MOE-430A)测定了小鼠树突状细胞瘤不同肺转移潜能的DD1和DG6细胞株的基因表达谱,同时我们从NCBI的基因表达数据库GEO和EBI的芯片数据库ArrayExpress at database下载了与树突状细胞(DC)发育相关的基因表达数据,下载的这些数据均是Affymetrix小鼠MG-U74av2芯片数据,我们用Affymetrix提供的两种芯片探针的匹配度筛选出最佳匹配的探针集,用内源的11个管家基因对2个自已测定的和32个从数据库下载的与DC发育相关的表达数据标准化,采用分层聚类分析、主成分分析、最近邻分析和自组织映射分析方法,将小鼠DD1和DG6细胞株的基因表达谱与数据库中树突状细胞(DC)发育各阶段的基因表达谱进行比较,结果发现小鼠树突状细胞瘤的基因表达谱与骨髓来源的造血干细胞和DC前体细胞相似,而与脾脏来源的相对成熟的DC差异较大,我们的结果支持肿瘤细胞来源于干细胞,干细胞发育障碍而导致肿瘤发生的理论,另外,结果还显示DD1和DG6基因表达谱与已知DC前体细胞的相似性,这个结果也提示合作单位分离的细胞为小鼠树突状细胞。4.含有同源异型结构的转录因子的生物信息学分析Affymetrix高密度表达芯片检测结果显示,高转移的DD1细胞株高表达同源异型基因En1(Engrailed 1 homeobox),低转移的DG6细胞株高表达lhx2(LIM homeobox protein 2)和prrx2(paired related homeobox 2)同源异型基因,提示同源异型基因与小鼠树突状细胞瘤可能有相关性。含有同源异型结构(homeobox)的蛋白在胚胎发育、基因表达调节、细胞分化、神经发生等方面发挥重要作用。同源异型结构的氨基酸残基在疏水性和带正电的氨基酸残基具有保守性,含有同源异型结构的蛋白质与基因启动子区富含A和T核酸序列结合。同源异型框与其它结构域同时存在,如PAX、POU、LIM、OAR、CUT、ELK、bZIP、SIX、PHD-finger、Engrailed等,近来还发现它通过基因融合或基因失调控方式参与肿瘤的发生,而且Nanog homeobox和OCT4(为POU复合homeobox结构)是胚胎干细胞维持自我更新过程所必需的。5.利用启动子区序列的同源性,寻找转录因子结合位点胰岛素启动因子1(IPF1,insulin promoter factor-1或PDX1,pancreatic duodenal homeobox-1)是含有同源异型结构(homeobox)蛋白质,和神经分化因子1(NDF1,neurogenic differentiation factor 1或NeuroD)、肝细胞核因子4(HNF4,hepatocyte nuclear factor 4)共同参与胰岛素基因表达的调控,均为MODY (maturity onset diabetes of the young)基因,这几个因子的功能缺陷是二型糖尿病发生重要的遗传因素。我们通过比较人、小鼠、大鼠这三种核蛋白氨基酸一级序列和结构域,发现其序列和结构十分相似,推测这些DNA结合蛋白与胰岛素基因结合的核苷酸序列相似;用ClustalW比较三者Promotor区的核苷酸序列,显示有几段核苷酸序列较为相似,同时搜索TRANSFAC基因转录数据库中NDF1、IPF1和HNF4蛋白核苷酸结合位点,发现核酸比对保守的部分序列与TRANSFAC数据库中这三个转录因子的DNA结合位点一致,另外一些核酸保守序列可能为其他未知DNA结合蛋白的结合位点。这为分子生物学实验寻找和验证新的胰岛素DNA结合蛋白与核苷酸的结合位点提供了基础。更多还原

【Abstract】 It is a common feature that the tumor cells metastasize to other tissues or organs, and the metastasis is the principal reason leading to death in individuals with tumor. Yet its molecular basis is not well understood. Especially, the recent studies have indicated that the key genes associated with metastasis and the mechanisms are diverse among the different tissues or of tumor types. Metastasis is a complex process involving multiple stages, such as local invasion, transit through the bloodstream, extravasation, proliferation and survival within a favorable target organ. We use the model of highly and poorly metastatic cell lines from the same parent (mouse Dendritic Cell Sarcoma, DCS) to examine differences of their gene expression profiles with Affymetrix high-density microarrays. We try to screen the genes related with lung metastasis of mouse DCS by the high-throughput method. The followings are the main points of our studies.1. The study of the genes associated with metastasis of mouse DCSMouse dendritic cell sarcoma (DCS) was a cell line that established and identified by Institute of Basic Medical Sciences at Chinese Academy of Medical Sciences. They isolated the highly and poorly metastatic cells (DD1 and DG6) from the parent cell line DCS by limited dilution method. It is a good model to study the metastasis of mouse DCS to lung for the nearly same genetic background between DD1 and DG6 subclones. We use the Affymetrix high-density microarrays to screen the genes that may be associated with metastasis by comparing the differences of genome-wide expression.The results show that expressed genes are quite different in transcription factors, RNA process, cell proliferation and apoptosis, transport, protein synthesis and modification, and other aspects between the DD1 (highly metastasis to lung) and the DG6 (poorly metastasis to lung) cells. Furthermore, several expression-difference genes have identified by the methods of RT-PCR and Immunocytochemistry, such as the transcription factor Ehf that influences the cell differentiation, potential tumor suppressors and apoptosis related gene Pedf, adhesion molecule Nectin-3 and the chemokine Mcp-1(CCL2). The results suggest these genes might be associated with the DCS metastasis.2. DNA methylation and the gene expression of mouse DCSWe choose some obvious up or down-regulated genes of transcription factors Smcy and Ehf, potential apoptosis molecule Pedf and adhesion molecule Nectin-3 to detect the DNA methylation status. It is to find whether the DNA methylation status in promoter region controls these genes over or lower-expression and try to explore the possible mechanism of gene expressions associated lung metastasis of mouse DCS. The genome DNA of DD1 and DG6 cells are transformed by bisulfite and as a template for PCR amplification using the primers that sequences not contain CG sequences. The DNA methylation status can be obtained by sequencing for conversion of the unmethylated cytosine is transformed into thymine with regions of interest. The result demonstrates that there is a site of CG difference in Pedf gene between DD1 and DG6 cells (changing C into T), and not obvious alteration in Smcy and Ehf genes between the two cells. From the results, we tendencious consider that the DNA methylation is not the main element to influence the three gene expression. There may be some other elements, such as the interaction with other transcription factors, to impact on the up or down-regulation expression.3. Gene expression analyses of mouse DCSWe isolated and screened two tumor cell clones DD1 and DG6 with different capacity of metastasis from the same parent cell line, a mouse dendritic cell (DC) sarcoma, using limited dilution method. The genome-wide expressions of DD1 and DG6 cells are done by Affymetrix’s MOE-430A microarray. The expressionprofiles related with mouse DC development are downloaded from GEO at NCBI and ArrayExpress at EBI database. In order to compare expression of DC sarcoma and DC developmental arrays which performed by MG-U74av2, we have screened the best matched probesets between MOE-430A and MG-U74av2 according to the probe identities from Affymetrix technical annotation. After the normalization of 11 housekeeping genes across the 34 arrays (2 DC sarcomas and 32 DC developmental arrays), all these expression profiles are analyzed by the methods of hierarchical clustering, principal component analysis, nearest-neighborhood and self-organizing maps. The results indicate that expression profiles of DC sarcoma are closer to the DC progenitors and hematopoietic stem cells from bone marrow compared with the sorted DCs from spleen. The results support the hypothesis that cancers (tumors or sarcomas) arise from stem cells. Furthermore, the results also suggest cells isolated from mouse are DC sarcoma cell lines for they are similar to the DC progenitors in gene expressions.4. Bioinformatics analysis on the transcription factors containing homeoboxThe results of Affymetrix microarray show that En1(Engrailed 1 homeobox)and Hlxb9(homeobox gene HB9)are up-regulated in DD1 cell, lhx2(LIM homeobox protein 2)and prrx2(paired related homeobox 2)are down-regulated in the cell. The result indicates that the homeobox gene may be related with metastasis of mouse DCS. The proteins with ’homeobox’ (HOX) domain play an important role in embryonic development, gene regulation, cell differentiation and tumorigenecity. Hydrophobic and positively charged residues in HOX domain are conservative, and the proteins with HOX bind to the rich A or T nucleotides in the gene promoters. Some other domains are found in the proteins together with HOX, such as PAX, POU, LIM, OAR, ELK, SIX, bZIP, PHD-finger, Engrailed, hexapeptide domains. These“complex homeobox”also take part in tumorigenesis by the way of gene fusions and deregulation. Recently, the new homeobox gene Nanog and OCT4(POU complex homeobox)are discovered that they are essential in self-renewal of embryonic stem cell.5. Finding binding sites by comparing the homology of promoter sequencesInsulin promoter factor-1(IPF1 or PDX1,pancreatic duodenal homeobox-1)is an important transcription factor containing homeodomain which takes part in regulating the expression of insulin gene as well as neurogenic differentiation factor 1(NDF1 or NeuroD) and hepatocyte nuclear factor 4(HNF4). All of them are MODY (maturity onset diabetes of the young) genes. The dysfunction of these factors is the key genetic element for type 2 diabetes. The sequences and domains of the three factors are homologous by the residues alignment of human, mouse and rat. It is supposed that the nucleotide binding sites are similar in the promoters of insulin gene. The ClustalW program of multiple alignment is used to compare the promoters of insulin genes. Meanwhile, the nucleotide binding sites of NDF1, IPF1 and HNF4 are searched in the TRANSFAC gene regulation database. The conservative nucleotide sequences are found in insulin promoters and some sequences are identified as the binding sites of several known transcription factors. There are also other conservative DNA sequences in the promoters perhaps are the binding sites of unknown factors. The alignment could offer a simple and applicable method to find and confirm the nucleotide binding sites of DBPs for the molecular biology experiments.更多还原