【作者】 尚小云；

【导师】 吴玉章；

【作者基本信息】 第三军医大学 ， 免疫学， 2009， 博士

【摘要】 细胞毒性T淋巴细胞(cytotoxic T lymphocytes,CTLs)在机体控制肿瘤中起重要作用,T细胞识别的是由抗原提呈细胞(antigen presenting cell, APC)表面MHC分子提呈给T细胞抗原受体(T cell antigen receptor, TCR)的一段多肽,即表位(epitope),基于CTL表位的治疗性肿瘤疫苗已经成为肿瘤综合生物治疗的重要策略之一。迄今为止已从60多种人肿瘤抗原中鉴定出170多个CTL表位,其中基于部分表位的肽疫苗已进入临床试验。临床试验结果表明:一方面,肽疫苗对于机体相对安全,而且易于大规模制备、纯化和质控,显示出诱人的发展空间;但另一方面,基于天然表位(wild type, WT)的肽疫苗免疫原性弱,很难在体内诱导出有效的抗瘤CTL反应。这是因为肿瘤在体内诱导了免疫系统对肿瘤抗原的免疫耐受,表现出免疫系统对肿瘤抗原的特异性免疫低应答或无应答。肿瘤抗原引起的免疫耐受有多种机制,总体可以分为中枢耐受和外周耐受两大类,这两类耐受所包含的机制也不尽相同。肿瘤抗原引起的免疫,既有中枢耐受也有外周耐受,若要在肿瘤免疫治疗中获得理想的疗效,就需要打破由肿瘤抗原引起的免疫耐受。打破免疫耐受最有效的方法就是通过改变致耐受抗原的分子结构,将这些经过改造的抗原给予机体,可特异性终止已建立的耐受。因此,如何对肿瘤抗原进行设计和改造以打破机体的免疫耐受,成了肿瘤治疗性肽疫苗研制的关键问题。在众多增强CTL表位肽免疫原性的策略中,对天然表位肽进行分子改造和修饰被认为是最有前景的方法之一。通过对天然表位肽进行单个氨基酸替换、多肽末端化学修饰或加入非天然氨基酸等可有效提高其免疫原性,诱导更强的CTL活性。诱发有效细胞免疫应答的基础就是T细胞与APC间必须形成稳定的TCR/肽-MHC三分子复合物(TCR/ peptide-MHC complex,简写为TCR /pMHC)结构,CTL表位肽通过两端的残基与MHC分子表面凹槽形成稳定的pMHC复合物,这些残基被称为表位锚着残基(anchor residues),表位与TCR结合的位点即为非锚着残基。肿瘤抗原的表位改造可以通过改善其与MHC的结合和稳定性达到增强肽免疫原性的目的,或改造TCR结合位点以改变TCR与pMHC结合能力,以增强T细胞的活化,从而克服T细胞耐受,达到使肿瘤消褪的目的。已有研究证实了基于锚着残基改造的候选肽在体内外的实验中均可以一定程度地增强肿瘤特异性CTL的增殖,但这类激动肽(agonist peptide)在肿瘤免疫治疗中都不能产生明显的抑瘤效果。本课题研究中,我们运用分子模拟技术对肿瘤抗原NY-ESO-1的HLA-A*0201限制性T细胞天然表位(wild type, WT)NY-ESO-1157-165进行了非锚着残基的替换。首先借助蛋白质结构数据库(protein database, PDB)数据库中WT特异性TCR-pMHC三元体晶体1G4-9C-A2 (PBD ID:2bnr),在Insight II工作站上建立1G4-9C-A2结构模型,借助分子动力学模拟、分子柔性对接等技术分析TCR分子与pMHC结合特征,在此基础上,通过计算机丙氨酸突变扫描方法分析天然表位中与TCR分子相互作用的关键位点。结合计算机丙氨酸突变扫描结果和TCR/pMHC相互作用的结构特征,我们以该表位的第四位和第五位为研究对象,对这两个位点进行了天然氨基酸的随机替换,借助结合自由能等计算方法筛选获得系列侯选APL。然后合成多肽通过体外和体内免疫学效应检测进一步筛选可以上调免疫应答的候选APL。体外免疫学效应主要包括肽-MHC分子亲和力检测,APL诱导的特异性CTL分泌细胞因子IFN-γ水平的检测,以及特异性CTL细胞杀伤实验,通过上述实验,我们筛选出了可以上调免疫应答并能和天然表位发生交叉反应的候选APL,该候选APL(NY-ESO-1157-165W5F,简写为W5F)是将天然肽第五位的色氨酸(tryptophan, Trp)替换为苯丙氨酸(Phenylalanine, Phe)。由于此天然肽NY-ESO-1157-165羧基端的半胱氨酸(cysteine, Cys)残基易被氧化,使短肽聚合形成二聚体,影响其免疫性。有研究将其羧基端的半胱氨酸替换为缬氨酸(valine, Val)后可以有效的增强其稳定性,并可以显著提高其免疫原性,在此设计基础之上,结合我们筛选的基于TCR结合位点改造的APL,我们尝试了同时将锚着残基和TCR结合位点进行替换,我们将第五位的色氨酸替换为苯丙氨酸,将第九位的半胱氨酸替换为缬氨酸,基于此,我们得到了一个新的APL:NY-ESO-1157-1655F9V(简写为5F9V)。我们以NY-ESO-1157-1659V(简写为9V)肽作为对照,分析了该APL与HLA-A2分子的亲和力,并结合临床病例,筛选了HLA-A2+NY-ESO-1+的食管癌病人,分离患者外周血单个核淋巴细胞(Peripheral Blood Mononuclear Cell, PBMC),分别用9V和5F9V刺激病人PBMC诱导特异性CTL反应,随后检测了APL诱导的特异性CTL分泌细胞因子IFN-γ的水平以及CFSE标记特异性CTL增殖情况,并结合pentamer技术检测诱导出的特异性CTL频率。结果显示,在PBMC可检测到特异性CTL的患者中,其PBMC分别经5F9V和9V肽刺激活化后,5F9V诱导的特异性CTL的频率和分泌细胞因子水平显著高于9V诱导的特异性CTL,其增殖能力也要强于9V肽诱导的CTL,对pentamer标记后的荧光强度进行分析,结果初步提示5F9V比9V诱导的特异性CTL具有更高比例的高亲合力CTL。随后,从特异性CLT库(repertoire)的角度探讨了激动肽5F9V提高天然表位免疫原性的机制,我们利用TCR Vβ各家族特异性抗体和TCR CDR3 spectratyping技术比较分析病人诱导前、天然肽与APL诱导后的特异性CTL库,研究发现APL诱导前后其特异性CTL库谱发生了偏移(bias),APL与天然肽诱导的特异性CTL库之间也发生了偏移,由此可以推测,5F9V增强的免疫学效应可能是诱发了一群新的具有高亲和力并能与WT肽发生交叉反应的特异性CTL。本研究将反向疫苗学技术与计算机辅助疫苗设计技术相结合,建立了基于分子模拟、分子动力学和结合自由能计算的计算机辅助疫苗设计技术平台,大大的提高了疫苗开发的效率。我们应用该平台对肿瘤抗原表位NY-ESO-1157-165进行了非锚着残基的替换,该疫苗打破了传统的锚着残基改造策略,将肿瘤治疗性多肽疫苗的设计思路由肽-MHC分子相互作用拓展至TCR与pMHC分子复合物的相互作用,我们首次将锚着残基和非锚着残基同时进行了替换。通过In silico分析结合实验研究发现,与天然表位相比,5F9V能形成更稳定的肽-MHC复合物,以及更稳定的TCR/pMHC相互作用。在体内和体外免疫学效应研究中,从细胞因子分泌、细胞杀伤功能及细胞增殖能力等多方面证实了5F9V肽能诱发比天然肽更强的CTL反应。最后通过对激动肽5F9V和WT肽特异性CTL的TCR库谱进行了检测,从分子水平深入分析,证实了APL引起交叉识别及打破免疫耐受的机制可能与其活化了一群新的具有高亲和力的特异性CTL有关,为肿瘤治疗性多肽疫苗设计提供了新的理论基础和技术路线。更多还原

【Abstract】 Induction of antigen-specific cytotoxic T lymphocytes (CTL) by therapeutic peptide vaccination is a promising approach for cancer immunotherapy. The specific cellular immune response starts from recognition by TCR of an immunogenic epitope presented in the context of the class I major histocompatibility complex (MHC-I) molecules. Thus, modulation of CTL response by manipulating T cell epitopes is a particularly attractive approach for cancer immunotherapy, because peptides from cancer cells are usually poorly immunogenic and often induce immune-tolerance. Vaccination with altered peptide ligands (APLs), which can be generated by appropriate amino acid substitutions at certain T cell epitopes, has become an attractive strategy to enhance specific T cell responses to tumors. This strategy can be achieved by two general approaches: 1) by increasing the affinity between the epitope and the MHC through substitution in the MHC anchor residues; or 2) by enhancing the interactions between the TCR and peptide-MHC (pMHC) complex through alteration at the TCR contact residues.Although APLs with altered MHC contact residues can efficiently activate tumor-specific T cells in vitro, vaccination with this kind of APLs has generally failed to elicit an effective anti-tumor CTL response that can lead to clinical tumor regression. APLs with increased pMHC complex affinity for the TCR molecule, which are designed by modifications at the TCR contact sites rather than the MHC anchor residues, have unexpected potency to induce stronger T cell responses and may even covert cross-reactive T cells from a tolerance state. According to the structure of the TCR/pMHC complex established by X-ray crystallography, recognition of an epitope by T cells is controlled by a few exposed TCR contact residues within the peptide. Several recent investigations have found that subtle changes at the TCR contact positions can dramatically alter the downstream signaling events that can lead to effects that can be in a range from induction of T cell anergy to enhancement of T cell functions, indicating that the analogue with substitution at TCR contact sites may provide considerable benefit in super-agonists or antagonist vaccine development. To date, there is no convenient method to guide the modification of TCR contact residues of T cell epitopes to change the affinity between pMHC and the TCR. In the past, only a few APLs were identified by methods such as eluting naturally occurring mutant peptides from tumor cells, high-throughput screening of synthetic combinatorial peptides libraries, and random phage displayed peptides libraries; however, these methods are costly and time consuming. It is thus necessary to develop a novel rational approach to guide such substitution.There have been many successful studies on evaluating the design of APLs for increased MHC binding affinity through the calculation of pMHC interaction energies using in silico techniques. We reasoned that calculation of the interaction energy between TCR and pMHC using computer-aided methods could be applied in the design of APLs for enhanced TCR engagement. However, the prediction of the interaction energy between TCR and pMHC is much more difficult than calculating pMHC interaction energies. Only a few theoretical approaches such as the free energy perturbation (FEP) method, regression method, and the statistical mechanics method have been developed to predict protein-protein binding affinity. In a FEP method-based study performed by Michielin and Karplus, the computed free energy difference in the binding of a particular TCR (A6) with a HLA-A2 restricted wild-type peptide (Tax) and a mutant peptide (Tax P6A) was shown in good agreement with the experimental value. Although this study has shed new light on the application of a molecular simulation approach to guide peptide modifications for alteration of TCR-ligand binding, the large computational burden made the application of the FEP method only applicable to established facility. In a recent study, Lai et al. applied a statistical mechanics method termed PMFScore, which is based on the potential of mean force (PMF), to calculate protein-protein interaction energies precisely and efficiently and here we aimed to test the feasibility of the application of PMF-based in silico approach in order to develop a more applicable approach for peptide modifications.The HLA-A*0201 restricted T cell epitope NY-ESO-1157–165 has been indicated as a promising candidate for T cell-based tumor vaccination strategies, however, several studies have revealed its defects in stability and bioavailability and its frequent failure to elicit robust anti-tumor CTL response. It has been demonstrated that a cysteine-to-valine substitution at position 9 in the NY-ESO-1 157–165 epitope can increase its immunogenicity due to markedly enhanced peptide binding to the MHC peptide binding grove. Although this MHC anchor residue substitution was also shown to possess slightly improved interactions with TCR than analogue peptides, the study did not specifically focus on the TCR contact residues.In this study, we generated an agonist analogue NY-ESO-1157–1655F9V with a tryptophane to phenylalanine substitution at TCR contact residue of NY-ESO-1157–165 based on a cysteine-to-valine substitution at position 9. This designed APL can elicit a stronger CTL response with cross-reactivity with the WT peptide. In conclusion, our findings demonstrated that the in silico method based on PMFScore could predict and guide T cell epitope modification of the TCR contact residues based on the structural information of TCR/pMHC triple complex. Our results provide important insights into the enhanced immunogenicity of epitopes through substitution at the TCR contact sites and revealed a novel molecular simulation approach for rational design of agonist peptides. It will be of interest to further examine the immunological effects of the 5F9V agonist in order to make this peptide to be applicable for antitumor vaccine design.更多还原