【作者】 应美丹；

【导师】 何俏军； 杨波；

【作者基本信息】 浙江大学 ， 药理学， 2011， 博士

【摘要】 背景：人间充质干细胞(hMSCs)在体内多种因素的调节下可以发育为成骨细胞,对正常的骨发育和维持骨代谢平衡十分重要。而另一方面,骨的发育过程异常与骨肉瘤的发生有密切关系,如果hMSCs在成骨分化的任意阶段发生恶性转化,那么就可能产生肿瘤干细胞(Cancer stem cells, CSCs)而导致骨肉瘤的发生。本论文就骨的正常发育机制和骨肉瘤发生的原因进行研究,试图寻找可以促进骨正常发育的新化合物,同时从CSC的视角分析骨肉瘤发生的机制,最后综合分析正常骨发育和肿瘤恶性转化过程中基因谱的差异,为进一步认识骨分化的分子机理提供实验基础和理论数据,为揭示骨肉瘤的发生机制及其治疗策略提供全新的思路。本研究将主要分为三个部分：(1)研究维甲酸及其信号通路对hMSCs成骨分化的调节作用及其分子机制,从分子水平了解正常骨发育的变化规律。(2)寻找骨肉瘤肿瘤起始细胞的表面标记物,并利用这些表面标记物分离和鉴定骨肉瘤的肿瘤起始细胞,探讨骨肉瘤发生的可能机制。(3)结合正常骨分化模型和骨肉瘤发生模型,利用基因芯片技术和生物信息学手段比较病理过程和正常分化过程中细胞的全基因谱的差异,寻找骨肉瘤发生早期的关键基因以及可以用于诊断和治疗骨肉瘤的潜在生物标记物,为临床上骨肉瘤的诊断和治疗提供新的靶点和思路。第一部分RARa信号通路调控人间充质干细胞向成骨细胞分化的作用及其机制研究目的：全反式维甲酸(ATRA)是维生素A的衍生物,机体内一定浓度的维甲酸可调节细胞的增殖、分化和成熟,是机体正常生长发育和各种生理活动必不可少的重要因子,同时维甲酸也是一种作用很强的分化诱导剂,在恶性肿瘤的发生与治疗方面也有着重要的作用。本课题主要研究ATRA对hMSCs成骨分化的调节作用及其可能的分子机制,以及低磷酸化状态RARa对hMSCs成骨分化的调节作用。方法和结果ATRA在hMSCs的成骨分化的早期过程中加入可以有效地促进碱性磷酸酶的活性,抑制hMSCs的细胞增殖,明显上调骨分化的主要转录因子Runx2的mRNA和蛋白水平,同时骨分化的标志蛋白OPN和Osteocalcin的表达也相应上调。利用基因芯片技术比较经ATRA诱导分化的hMSCs中全基因表达谱的变化,发现大量骨分化相关基因的表达均出现2倍以上的增加。随后我们在正常培养的hMSCs上检测到了内源性ATRA的分泌。利用慢病毒载体在hMSCs上高表达低磷酸化的RARa,培养7天后hMSCs会自发地出现碱性磷酸酶活性增加以及钙沉积增多,细胞增殖也明显受到抑制,同时骨分化的标志蛋白OPN和Osteocalcin的表达也被上调。另一方面,ATRA和经典的BDM诱导hMSCs成骨分化过程中,BMP8a的表达出现明显上调,而且高表达RARaS77A的hMSCs出BMP8a的表达也出现明显上调。在hMSCs过表达BMP8a以后,细胞自发地会出现钙沉积的增多和骨分化的标志蛋白OPN的上调。此外,Fgf8f在高表达RARaS77A的hMSCs中蛋白表达水平出现明显上调,而在hMSCs过表达Fgf8f以后,hMSCs也自发地会出现钙沉积的增多。结论：本研究证明ATRA和地塞米松联合在hMSCs成骨分化的早期应用可以明显地促进hMSCs的成骨分化,其中ATRA可能是通过降低RARa的磷酸化水平来激活下游BMP8a和Fgf8f信号通路,进而启动Runx2对下游骨分化靶基因ALP、OPN、Osteocalcin等的转录,最终促进了hMSCs的成骨分化过程。第二部分人骨肉瘤肿瘤起始细胞以及骨肉瘤肿瘤形成细胞的鉴定目的：骨肉瘤是一种起源于间叶组织的恶性肿瘤,目前也被认为是一种细胞分化异常的疾病。肿瘤干细胞理论可以很好的解释骨肉瘤对放化疗的抵抗、复发以及转移,因此阐明骨肉瘤的细胞起源,鉴定并分离出骨肉瘤肿瘤起始细胞对于骨肉瘤的发病机制、预防以及研究针对肿瘤干细胞的新型药物都有很重要的意义。本论文试图寻找骨肉瘤肿瘤起始细胞的表面标记物,并利用这些表面标记物分离骨肉瘤肿瘤起始细胞,揭示骨肉瘤肿瘤起始细胞的生物学本质,从而将骨肉瘤肿瘤起始细胞(Osteosarcoma-initiating cells, OSICs)和骨肉瘤肿瘤形成细胞(Osteosarcoma-forming cells, OSFCs)加以区分。方法和结果：(1)利用连续接种裸小鼠的方法可以提高OSICs的自我更新能力进而产生更多的子代细胞促进肿瘤组织的形成,本研究发现CD49f+细胞可以维持骨肉瘤肿瘤的生长。通过对CD49f+细胞及其亲代细胞进行进一步的体内外谱系追踪实验,研究发现CD49f+细胞是由CD49f-细胞产生的。进一步比较CD49f+和CD49f-细胞的生物学特性,研究发现CD49f+细胞在成瘤能力和化疗药物耐受方面均不如亲代CD49f-细胞,更为重要的是, CD49f-细胞还可以产生CD133+细胞。在30.5%的CD49f UT2细胞中仅含有0.97%的CD49fCD133+细胞,而这少数的CD49fCD133+细胞亚群却比CD49f-细胞可以产生更多的CD49f+细胞亚群。在可以生成肿瘤的CD49fCD133+、CD49f、CD49f+、CD133+和CD133-所有亚群中,只有CD49fCD133+细胞表现出最高的成瘤率(75%), CD49f、CD49f+、CD133+和CD133-分别为55.6%、44.4%、33.3%和25%。(2)裸小鼠移植瘤实验结果显示,尽管CD49f、CD49f、CD133+和CD133-细胞亚群均表现出一定的体内成瘤能力,但是OSICs中的CD49fCD133+细胞亚群比上述4个亚群具有更强的成瘤能力。此外,在连续的移植瘤模型中子代CD49f+细胞尽管成瘤能力不强,但是却具有更强的成骨分化能力。相应地,CD49fCD133+细胞具有和CD49f-、CD49f、CD133+、CD 133-、CD49fCD133细胞类似的脂肪细胞分化能力和较高的成瘤率,但是CD49fCD133+细胞的成骨分化能力却相对下降了。在UT2细胞和U20S细胞上也观察到了同样的现象,即具有较强成瘤能力的UT2细胞的成骨分化相关基因发生了下调,而脂肪分化相关基因没有进一步下调。这些数据说明CD49fCD 133+田胞以及经连续移植瘤实验提高了自我更新和成瘤能力的UT2细胞均伴随着成骨分化能力减弱的现象。(3)为了明确CD1 17+和Stro-1+是否和CD49fCDl33+细胞亚群具有类似的自我更新和成瘤能力,本研究在具有较高成瘤率的UT2细胞上分析这些标记物的成瘤能力,将500个CD117+、CD117-、Stro-1+、Stro-l-、CD49f、CD49f+、CDl33+和CD133-UT2细胞分别注射到NOD/SClD小鼠背部,研究显示和CD49f-、CD49F+、CD133+以及CD133-UT2细胞相比,CD117+、CD117-、Stro1+、Stro-l-UT2细胞的成瘤率很低或者没有。(4)目前,关于CD49f和CD133+对于OSICs细胞的维持和在干细胞标记物中的作用仍然不是很清楚,基因芯片数据显示Wnt、Notch、BMI1和TGF-β信号通路在经移植瘤模型激活的UT2细胞中明显上调,这就提示这些通路在UT2细胞中的CD49fCD 133+细胞亚群的自我更新和成瘤能力中十分重要。结论：本研究首次发现了OSICs中富集着CD49fCD133+细胞,而且OSICs产生的CD49fCD 133+细胞存在着成瘤能力和成骨分化能力互斥的关系,这些研究不仅为进一步寻找和CD49fCD 133+联合的其他肿瘤干细胞潜在标记物以及研究不同信号通路在OSICs和OSFCs中的作用奠定了前期基础,也证明在区分CSCs.子代祖细胞和肿瘤形成细胞中联合应用谱系追踪、成瘤能力对比、药敏实验等多种手段是非常必要的。第三部分调控成骨分化和骨肉瘤发生的差异基因表达谱研究目的：随着我们对干细胞认识的加深,越来越多的证据提示骨肉瘤应该是由于hMSCs在成骨分化过程中发生基因以及表型的改变终止了正常分化进程而形成的分化疾病。因此本部分的研究将第一部分ATRA促进hMSCs的正常成骨分化模型和前期的ATRA诱导骨肉瘤U20S细胞的分化治疗模型,以及第二部分骨肉瘤细胞模型,利用人类基因组表达谱芯片和生物信息学手段对调控成骨分化和骨肉瘤生成的基因进行了严格比较和筛选,全面分析和比较病理过程和分化过程中全基因谱的差异,试图从中寻找骨肉瘤发生最早期及最关键的因子,并借此寻找潜在的诊断和治疗骨肉瘤的生物标记物,为将来治疗骨肉瘤提供一些新的靶点和思路。方法和结果：(1)通过分析4株具有不同遗传背景的骨肉瘤细胞的全基因谱表达,发现了438个基因的表达在4株细胞上呈现共同的变化趋势,其中共同表达上调2倍以上的基因有133个,而共同表达下调2倍以上的基因有305个,通过IPA软件分析这些基因和骨肉瘤生成密切相关；(2)通过分析hMSCs的骨分化模型、骨肉瘤U20S细胞的分化治疗模型和4株具有不同遗传背景的骨肉瘤细胞的全基因谱表达情况,研究发现了295个在骨分化和骨肉瘤生成中分别具有相反的表达趋势的基因,通过DAVID软件对其生物学功能进行归类分析发现这295个差异基因主要为细胞接触、跨膜蛋白、转录因子、激酶、泛素样蛋白、胶原、代谢酶、受体、分泌的细胞外信号、骨系统发育蛋白、蛋白酶等；(3)我们在295个差异基因中发现了13个转录因子,分别为EPAS1、FOS、GTF2A1L、MXD1、CDKN2A、CDKN2B、CHURC1、GLIS3、NFKBIZ、HOXB6、FOXF1、BCL11A和EGR2。其中EPAS1、FOS和GTF2A1L这3个基因表达水平在骨形成过程中上调,但是在骨肉瘤发生过程中却明显下调,而MXD1、CDKN2A、CDKN2B. CHURc1 GLIS3和NFKBIZ这6个基因在骨肉瘤发生过程中明显下调,但是在骨形成过程中变化不明显,只有HOXB6、FOXF1、BCL11A和EGR2这4个转录因子在骨肉瘤发生过程中发生了明显的上调,但是在骨形成过程中变化不明显的。(4)我们在上述的295个基因中发现了包括ADAMTS2、ADRB2、B4GALT1、BTN3A1、CD14、CDKN2A、CDKN2B、CH13L1、COLLA1、CYPLB1、DPP4、DPYSL3、ENG、EPAS1、FN1、FOS、HOXB6、IGFBP3、ITGA5、MCM7、OPTN、RARRES1 SNAP25和SPARC在内的24个潜在的生物标记物的变化在骨肉瘤发生过程和骨形成过程出现明显差异,其中ADRB2、HOXB6和MCM7这3个潜在的骨肉瘤标记物,对于寻找骨肉瘤发生最早期及最关键的步骤,以及寻找潜在的诊断和治疗骨肉瘤的生物标记物具有重要的指导意义。结论：本部分研究报道了295个在骨分化和骨肉瘤生成中分别具有相反的表达趋势的基因,而且发现了HOXB6、FOXF1、BCL11A和EGR2这4个转录因子与骨肉瘤发生密切相关,特别是寻找到了ADRB2、HOXB6和MCM7这3个潜在的骨肉瘤标记物。上述结果对于寻找骨肉瘤发生最早期及最关键的因素,以及寻找潜在的诊断和治疗骨肉瘤的生物标记物具有重要的指导意义。更多还原

【Abstract】 IntroductionHuman mesenchymal stem cells (hMSCs) are thought to be multipotent cells, which are present in adult marrow, that can replicate as undifferentiated cells and that have the potential to differentiate to osteoblast cells and contribute to bone development and fracture repair. Meanwhile, osteosarcoma transformation may be associated with osteoblastic differentiation of hMSCs, because osteosarcoma may be caused by genetic and epigenetic disruptions of osteoblastic differentiation pathway in hMSCs. Thus, this study focus on the cellular and molecular pathways involved in osteoblastic differentiation and osteosarcoma pathogenesis, investigated the role of ATRA in control of osteoblastic development of hMSCs, characterized the sub-population of cells within osteosarcoma possess cancer initiating properties, and analyzed potential links between defective osteogenic differentiation and bone tumorigenesis. Understanding the molecular pathogenesis of human osteosarcoma could ultimately lead to the development of diagnostic and prognostic markers, as well as targeted therapeutics for osteosarcoma patients.The current study include three parts:part 1 to evaluate the role of ATRA and its pathway in controlling of osteoblastic development of hMSCs; part 2 to identify and characterize cancer stem cells within osteosarcoma; part3 to analyzed potential links between defective osteogenic differentiation and osteosarcoma transformation using microarry analysis and Bioinformatics.Part 1 RARa-Bmp8-Fgf8 signaling in mediating osteoblastic lineage development of human mesenchymal stem cellsObjective:Retinoic acid (RA) and their receptors are present in differentiating tissues during development and are thought to play an important role in regulating various processes such as proliferation, differentiation and maturation of cells. ATRA is also known as the most potent differentiation inducer to treat malignant disease. We asked in this study whether retinoid signaling could have a role in the differentiation of human mesenchymal stem cells (hMSCs) to an osteoblastic lineage.Methods and results:ATRA effectively enhances the osteoblast-specific alkaline phosphatase (ALP) activities in both hMSCs and the supernatant and inhibits proliferation of early stage of hMSCs. The expression of osteoblastic differentiation marker genes, including OPN, osteocalcin, and RUNX2 are also up regulated in the presence of ATRA. In addition, DNA microarray analyses show that ATRA up regulate expression of multiply osteoblastic differentiation marker genes in hMSCs. Transduction of RARaS77A, a phosphorylation-defective form of RARa, mimics the effect of ATRA in mediating osteoblastic differentiation of hMSCs. Transduced hMSCs automatically activate ALP and enhance calcium deposition. RARαS77A also inhibit hMSCs’proliferation and induce OPN, Osteocalcin expression. Interestly, up-regulation of BMP8a is observered in ATRA, BDM and RARaS77A induced hMSCs osteoblastic differentiation. Overexpression of BMP8a in hMSCs shows calcium deposition and OPN up-regulation. Besides BMP8a, FGF8f is also up-regulated in RARaS77A transduced hMSCs, and FGF8f overexpression in hMSCs will cause calcium deposition. Conclusion:ATRA combine dexamethasone effectively enhance osteoblastic early differentiation of hMSCs. ATRA-induced RARa hypophosphorylation is involved in mediating osteoblastic lineage development of hMSC. BMP8a and FGF8f induction activated by ATRA or RARa hypophosphorylation are involved in regulation of the earlier and later stage of osteoblastic differentiation, respectively.Part 2 Identification of cells initiating human osteosarcoma and generating osteosarcoma-formation cellsObjective:Osteosarcoma, generally considered a differentiation disease, arises around the growth plate of long bones. Identification and understanding of OSIC are crucial to develop novel cancer stem cell-directed treatments against osteosarcoma, which should reduce therapy resistance, relapse, and the toxicity associated with current, non-selective chemotherapy in the treatment of osteosarcoma. In this study, we identified stem cell biomarkers of cancer stem cells within osteosarcoma, and isolated and characterized osteosarcoma initiating cells and osteosarcoma tumor forming cells.Methods and results:(1) Based on the idea that enhanced self-renewal of OSIC by serial transplantation will lead to an increasing generation of more differentiated progenies to accelerate tumor formation, we identify the progressively increased CD49f+ daughter cells that are responsible to constitute the bulk of osteosarcoma. By further seeking the lineage relationship between increasingly expanded CD49f+ progenies and their parental cells using lineage-tracking strategies both in vitro and in vivo, we determine that CD49f+cells are generated by CD49f-subpopulation. Those CD49f+ cells are weaker in tumorigenicity and drug resistance compared to their parental CD49f-subpopulation. Importantly, these CD49f-cells also generate CD 133+subpopulation. Among 30.47%CD49f_-subpopulation in UT2 cells, there are only 0.97%CD49f-CD133+cells. Such rare CD49f-CD133+subpopulation has higher capacity than CD49f-cells in generation of CD49f- cells. Whereas all those CD49f-CD133+, CD49f-, CD49f+,CD133+, and CD133-subpopulations can generate in vivo osteosarcoma formation, CD49f-CD133+cells exhibit the highest capacity of tumorigenicity (75%) than CD49f-(56%), CD133+(42%), CD49f+ (44%), CD133-(33%), or CD49f-CD133- (0%) subpopulation, respectively. (2) Our studies demonstrate that whereas all CD49f-, CD49f+, CD133+, or CD133- subpopulation show in vivo tumorigenicity, OSIC-enriched CD49f-CD133+cells possesses higher tumorigenicity than its direct progenies,CD49f+ cells, as well as other single subpopulations. Moreover, during the processes from progressively enhanced OSIC self-renewal by serial transplantation to increasingly production of their progenies forming the bulk of osteosarcoma, we observe that those CD49f+ daughter cells possess lower capacity of tumorigenicity but gain more osteogenic feature that can be induced by osteogenic reagent, compared to their parental cells. Correspondingly, while OSIC-enriched CD49f-CD133+cells retain similar capacity of adipocyte fate compared to CD49f-, CD49f+, CD133+, CD49f-, or CD49f-CD133- subset, their high tumorigenicity is associated with decreased osteogenic differentiation. These dynamic changes are also observed in the developmental processes from U2OS to UT2 cells induced by serial transplantation, in which high tumorigenicity of UT2 cells is associated with an inhibition in osteogenic differentiation but not in adipocyte fate, as shown by that there is a decreased expression of osteoblast genes than adipocyte genes in UT2 cells compared to U2OS cells. These data demonstrate that in CD49f-CD133+ subpopulation, the progressively increased activities of self-renewal and tumorigenicity from U2OS to UT2 cells by serial transplantation are associated with inhibited osteogenic fate. (3) To check whether CD117+and/or Stro-1+subpopulation share in vivo self-renewal and tumorigenicity with CD49f-CD133+subpopulation, we chose non-transformed UT2 cells first for xenotransplantion analysis. The same amount of 500 cells isolated by using CD117+, CD 117-, Stro-1+, Stro-1-, CD133+, CD 133-, CD49f-, or CD49f+ antibodies was engrafted into NOD/SCID mice. We found that CD117+, CD117-, Stro-1+, or Stro-1-cells show no or low tumorigenicity compared to CD133+, CD133-, CD49f-, or CD49f+cells. (4) Currently, the functions of CD49f- and/or CD133+in the properties of OSIC remain unknown beyond their roles in serving as selection markers. However, our initial data indicate that in UT2 cells generated by serial transplantation, the molecules that are crucial in mediating Wnt, Notch, and TGF-(3 signaling pathways have been significantly up-regulated compared to MSC and U2OS cells. This suggests that in UT2 cells, the enriched OSIC self-renewal and tumorigenicity represented by CD49f-CD133+subpopulation are related to activation of those pathways.Conclusion:In summary, we have identified that CD49f-CD133+subpopulation is enriched with OSIC. This OSIC-enriched CD49f-CD133+subset endures new phenotype of high tumorigenicity with inhibited osteogenic fate, compared to its progenies showing limited tumorigenicity with more differentiated osteogenic feature. Hence, these studies have not only laid down the foundation to further identify possible candidate partner(s) of CD49f-CD133+subpopulation and to determine the role of individual self-renewal pathways in OSIC vs. OSFC, but also provide the evidence for the first time that in the course of identifying cancer stem cells, a combined approaches of lineage tracking, tumorigenicity comparison, and differentiation evaluation are important to differentiate cancer stem cells from their progenies, cancer formation cells.Part 3 Study of gene expression profile regulating osteogenesis and osteosarcoma transformationObjective:With a rapid expansion of our knowledge about stem cell biology, emerging evidence suggests osteosarcoma should be regarded as a disease of cell differentiation caused by genetic and epigenetic changes that interrupt osteoblastic differentiation of human mesenchymal stem cells. We analyzed potential links between defective osteogenic differentiation and osteosarcoma transformation using microarry analysis and bioinformatics, trying to find the key step controlling osteosarcoma transformation and the potential diagnostic and prognostic markers for osteosarcoma.Methods and results:(1) Microarry was used to analysis the gene profiles of four osteosarcoma cells (TTC444, TTC606, UT2 and U2OS) vs. hMSCs.438 genes are sorted to show similar mRNA expression level change in these four cells with 133 genes up-regulated and 305 genes down-regulated; (2) A combination analysis of the gene profiles of ATRA-induced hMSCs differentiation, ATRA-induced U2OS differentiation, RARaS77A-induced U2OS differentiation, four osteosarcoma cells. The mRNA expression levels of 295 genes are inverse between osteogenesis and osteosarcoma transformation. These 295 genes are distributed in the protein family of cell junction, transmembrane region, transcriptional factor, kinase, ubiquitin like protein, collagen, metabolic enzyme, receptor, secreted extracellular signal, skeletal system development, Protease etc. (3) There are 13 transcriptional factors, including EPAS1, FOS, GTF2A1L, MXD1, CDKN2A, CDKN2B, CHURC1, GLIS3, NFKBIZ, HOXB6, FOXF1, BCL11A and EGR2. The mRNA level of EPAS1, FOS and GTF2A1L are up-regulated in osteosarcoma transformation and down-regulated in osteogenesis. The mRNA level of MXD1, CDKN2A, CDKN2B, CHURC1, GLIS3 and NFKBIZ are down-regulated in osteosarcoma transformation and stable in osteogenesis. Only HOXB6, FOXF1, BCL11A and EGR2 are up-regulated in osteosarcoma transformation and stable in osteogenesis. (4) ADAMTS2, ADRB2, B4GALT1, BTN3A1, CD14, CDKN2A, CDKN2B, CHI3L1, COL1A1, CYP1B1, DPP4, DPYSL3, ENG, EPAS1, FN1, FOS, HOXB6, IGFBP3, ITGA5, MCM7, OPTN, RARRES1, SNAP25 and SPARC are sorted as potential cancer biomarkers with inverse expression of mRNA between osteogenesis and osteosarcoma transformation. Only ADRB2, HOXB6 and MCM7 are recognized as best candidates for diagnostic and prognostic markers.Conclusion:It is concluded that 295 genes are inversely expressed in osteogenesis and osteosarcoma transformation. HOXB6, FOXF1, BCL11A and EGR2, as transcriptional factors, play important role in osteosarcoma transformation. Moreover, ADRB2, HOXB6 and MCM7 are powerful osteosarcoma diagnostic and prognostic markers. This study will provide insight into the pathogenesis of osteosarcoma, as well as targeted therapeutics for osteosarcoma patients.更多还原