肿瘤已成为全球首要致死原因，给人类社会带来了沉重的疾病负担。深入具体地认识肿瘤发生的分子机制，将极大地协助大家精准地治疗肿瘤，提高患者生活质量和生存时间。作为一种基因组不稳定的体细胞遗传病，肿瘤之间存在很大异质性，并且本身处于不断“进化”的状态，这些特性使得鉴定、验证更多肿瘤相关基因，并实现靶向治疗，依然是一件充满挑战的研究工作。

  本博士学位论文包括以下三项研究：

1、为发现更多肿瘤抑制基因，我们设计并开展了利用转座子在小鼠体内筛选肿瘤抑制基因的工作。在Blm缺失背景的小鼠体内，我们利用piggyBac转座子介导体细胞发生插入突变，干扰基因正常功能，从而诱发肿瘤生长。接着，我们收集了肿瘤标本，提取肿瘤基因组DNA，利用piggyBac序列标签克隆突变位点，并分析筛选的候选肿瘤基因。我们发现，小鼠体内高频次的转座可以导致小鼠胚胎致死；中等频率的转座会引起小鼠出生后生长发育异常；即使低频次的转座也会使小鼠整体生存周期缩短，并诱发肿瘤。在克隆piggyBac插入位点之后，我们可以看到，相比小鼠尾巴中转座子的整合，肿瘤标本中piggyBac插入位点存在明显的富集现象，提示肿瘤克隆扩增特性；piggyBac插入位点数量及其分布与已有研究比较一致；然而，在目前有限的肿瘤标本及测序分析中，我们尚未发现明确的肿瘤抑制基因。本课题的开展，提示我们利用piggyBac在体筛选肿瘤相关基因是可行；当然，为使研究更深入，限定筛选肿瘤的类型并完善相关专业条件也需要考虑。

2、为研究Cdk5rap3在小鼠肝细胞癌发生中的作用，我们在肝细胞特异的Cdk5rap3条件性敲除小鼠中，利用DEN诱发肝细胞癌的生长，通过与野生型小鼠观察比较肝癌发生率及肿瘤生长状态，明确该基因缺失对肝癌发生的影响。观察发现，肝细胞特异地敲除Cdk5rap3显著提高了小鼠肝癌发生率，其肿瘤生长数量和质量均远远高于野生型小鼠；让人意外的是，病理分析发现肿瘤细胞中Cdk5rap3表达明显升高，这些细胞并非来自基因敲除的肝细胞。这些结果提示我们，条件性敲除部分肝细胞中Cdk5rap3可以显著提高小鼠肝细胞癌易感性；这一表型的改变更多地来自肝脏内细胞与细胞或细胞与微环境之间的相互作用。

3、我们利用全外显子组测序发现在Flt3-ITD敲入小鼠中存在一个耐药变异（Flt3-ITD c.2076T>A）。为验证该变异会引起肿瘤细胞耐药，我们克隆了cDNA，并将其回复为野生型碱基（Flt3-ITD c.2076T），在分别转化了Ba/F3细胞系之后检测了细胞对Quazartinib（AC220）、Sorafenib和Ponatinib的敏感性。与预期一致，我们发现，两个编码序列均可顺利转化Ba/F3细胞，均显示了持续活化的激酶活性；与Flt3-ITD c.2076T相比，Flt3-ITD c.2076T>A使得肿瘤细胞对Quazartinib和Sorafenib表现出耐药，而两株细胞对Ponatinib均敏感，与已有报道结果一致。该实验证实了，Flt3-ITD敲入小鼠中发现的变异与人类白血病细胞FLT3中出现的Gate-keeper耐药突变一样，均会导致肿瘤细胞对Quazartinib和Sorafenib产生耐药性，而Ponatinib可以克服这种耐药。小鼠模型体内存在的耐药突变，提醒我们在使用小鼠模型开展实验研究时需要谨慎设计对照实验，以排除潜在的影响因素。

  总之，三方面的课题，分别从正向遗传学和反向遗传学的角度，使得自己可以利用多种遗传手段进行肿瘤学相关研究，并对它们有了一定的理解和掌握。

  Tumor has been the leading cause of the death all around the world and imposed an enormous burden on human society, of which the same situation China has to face. We believe there will be more chances to practice targeted therapies, to help relieve the pain of patients, and improve the overall survival in the near future based on the better understanding of the molecular mechanisms. Known as a disorder of somatic genetics with genomic instability, tumorigenesis is a long evolutionary process with multiple steps including mutations in cancer associated genes, clonal growth of abnormal cells and selection for the winner adapted to the environment. So it is essential to uncover the drivers and the pathways of the process. However, given the facts of the heterogeneity of tumors and the complexity of the genome, it is still a great challenge to decipher genomic information into general understanding of cancers, to explore the potential oncogenic function of candidate genes, and to develop more targeted drugs and methods.

  In this thesis, three cancer research projects were described:

1）To discover more potential tumor suppressor genes (TSG), we designed the project to screen for TSG using piggyBac transposon (PB) mediated insertional mutagenesis system in Blm-deficient mice, which elevates the rate of mitotic recombination by 20-fold. In the somatic cells of these mice, the normal transcription of the genes would be disrupted randomly when PB integrated into them correctly and the genes would lost function completely when loss-of-heterozygosity (LOH) happens. So tumors would be induced when enough TSGs had been trapped. Given the tag sequences of PB, we could clone the insertions sites from the cancer genomes and the common insertional sites (CIS) would be more likely to be cancer related. Here, we generated several PB transgenic mouse lines and found that the frequent transposition leads to the embryonic lethality; the moderate transposition caused developmental and growth abnormality; and even the low-frequency transposition effected the total survival of the mice compared to the control. We tried to analysis the insertional sites through Splinkerette-PCR combined with next generation sequencing and we found the reads number of some sites were enriched indicating clonal proliferation of the cells, which is consistent to other studies. So far, however, we could not find any confirmed TSGs, and maybe more samples needed to be analyzed. The Results we’ve got now told us it was possible to use piggyBac transposon system to find cancer associated genes without gain-of-function design while it would be much better to do the screening in a specific type of tumor.

2) To explore the potential role of Cdk5rap3 gene in Hepatocellular Carcinoma (HCC) and the underlying mechanism, we established a conditional knockout mouse (CKO) with specific deletion of Cdk5rap3 in hepatocyte and then we investigated the susceptibility to the DEN-induced HCC in these CKO mice. The results showed that partially deletion of Cdk5rap3 in mouse liver promoted HCC incidence and growth significantly. Unexpectedly, the expression of Cdk5rap3 in cancer cells were upregulated, which meant tumor cells originated from wild type cells rather than cells with Cdk5rap3 deprived. Occasionally, nodules with negative expression of Cdk5rap3 were found. Since the results and clues were limited, we were not sure the formation of the nodules, a primary tumor from liver cell without Cdk5rap3 or metastatic cells that has already been transformed and the gene was deleted after then. Anyway, the results indicated Cdk5rap3 played important role in the formation of HCC and had distinct functions at different phases and environments. In pre-cancer stage, the deletion of Cdk5rap3 in a small number of cells will disturb the balance of liver cells and created a microenvironment suitable for tumorigenesis; During the transformation, the cells will need the higher expression of Cdk5rap3, or instead of oncogenic function, the upregulation was only response to the transformation. In conclusion, the phenomenon observed is interesting and deserves further work.

3)Previously, we found a single nucleotide variation (SNV), called Flt3-ITD c.2076T>A, using whole exome sequencing (WES) in a Flt3-ITD knockin mouse model, which is well-known resistant mutation in AML patients. To confirm whether the SNV we discovered in the mice will bring the same drug resistance, we carried out the experiment. The Flt3-ITD cDNA with SNV was cloned and converted into wildtype nucleotide, named Flt3-ITD c.2076T. Then we transduced Ba/F3 cell line with these two versions of Flt3-ITD sequences and the cells were transformed successfully. At last we tested the sensitivity to the different tyrosine kinase inhibitor (TKI), Quazartinib (AC220), Sorafenib and Ponatinib. As we expected, there was a significant difference between the cell lines transduced with Flt3-ITD c.2076T>A and Flt3-ITD c.2076T, of which the former was resistant to Quazartinib and Sorafenib while both of them were sensitive to Ponatinib, which is consistent to the published result. Our result clearly shown us the SNV would give the leukemia cells the feature of drug resistance and the effect in the mice would be the same. So we need to pay attention when we use the model and explain the result of experiments as the SNV should not be ignored. But actually, this had given us a perfect model for drug resistance research in vivo. 

  All in all, the projects above focused on solve some practical problems in cancer research, using both forward and reverse genetic approaches. Through carrying out the experiment, I have learnt lots of different technologies and information.