【作者】 王伟力；

【导师】 尉军； Feng-Ming (Spring) Kong；

【作者基本信息】 吉林大学 ， 医学基因组学， 2014， 博士

【摘要】 肺癌的发病率和死亡率均逐年上升，虽然综合治疗在不断进步，其5年总生存率仍不足15%。临床上发现，病情相似的肺癌患者接受近乎相同的临床治疗后，其治疗疗效及副反应却差异较大，提示肺癌患者间存在较大的临床异质性。因此，区分不同的肺癌亚型而进行个体化治疗成为当今肺癌研究的热点问题之一。从而，发掘能够在治疗前评估预后、预测治疗疗效及毒性的生物标志物（Biomarker）成为当前肺癌个体化治疗研究的重要途径。胸部放射治疗是肺癌有效治疗手段之一，放射性肺损伤（Radiation-inducedlung injury, RILI）是肺癌胸部放疗最主要的并发症，限制了肺癌放疗的有效实施而影响放疗疗效。肺癌患者放疗后的总体预后及RILI的个体间差别较大。若在放疗前能够较准确评估肺癌患者的预后及疗效，预测重症RILI的发生，实施个体化的放疗方案，将有助于降低治疗相关死亡率，从而提高肺癌总生存。遗传学研究表明，单核苷酸多态性（Single nucleotide polymorphisms, SNPs）作为第三代遗传标记，代表着人类基因组中最丰富的遗传变异，其被作为候选的恶性肿瘤生物标志物而被研究多年。多个SNPs已被证实与肺癌的遗传易感性和/或治疗疗效及总体预后相关，其中DNA修复通路（DNArepair pathway）基因的SNPs作为肺癌生物标志物的研究报道最多。典型的范例是ERCC1基因多态性与肺癌的铂类化疗疗效相关。免疫学研究表明，外周血的肿瘤相关抗原（Tumor associate antigens, TAAs）/免疫调节分子（Immuno-regulatory molecules, IRMs）的自身抗体可在肿瘤形成实体肿块而被影像学检查发现前数月至数年即可被检测到。所以，TAAs/IRMs自身抗体作为肿瘤的生物标志物有其天然的优势。Early-CDT lung是第一个被研发出的诊断试剂盒用于肺癌的早期诊断。同时，许多研究也证实某些TAAs自身抗体与肺癌的预后及治疗疗效相关。在上述研究背景的基础上，本研究选择SNPs和TAAs/IRMs自身抗体作为候选的生物标志物。检测非小细胞肺癌（non-small cell lung cancer, NSCLC）患者个体间基因多态性及TAAs/IRMs自身抗体水平的差异，并与NSCLC胸部放疗后总生存及放射性肺炎做相关性分析。通过筛查阳性的生物标志物，尝试阐明NSCLC患者间的异质性。本研究结果显示：（1）遗传学方面：DNA修复基因的5个SNPs（ERCC2:rs238406、ERCC1:rs11615、ERCC1:rs3212948、XRCC4:rs9293329和XRCC4:rs2075685）与NSCLC患者总生存显著相关联。而且ERCC2:rs238406、ERCC1:rs11615与胸部放疗剂量对NSCLC患者总生存的影响存在交互作用。DNA修复基因的2个SNPs（XRCC4:rs1478486CC和XRCC4:rs2075685）与放射性肺炎（≥2级）显著相关联。（2）免疫学方面：在3个免疫调节分子及4个肿瘤相关抗原中未发现与NSCLC总生存相关的生物标志物。但发现CD25及OCT4II自身抗体水平与放射性肺炎（≥2级）显著相关联。本研究工作是NSCLC在免疫遗传学领域生物标志物研究的前沿课题，从遗传学和免疫学两个方面去筛查NSCLC胸部放疗后总生存及放射性肺炎相关的生物标志物。实验结果将有助于肺癌个体临床亚型的鉴别，为NSCLC个体化治疗临床策略的制定提供理论依据，最终将有助于提高NSCLC的总体临床疗效及总生存率。更多还原

【Abstract】 ObjectivesThoracic radiation therapy (TRT) is one of the most effective modalities innon-small cell lung cancer (NSCLC) management. However, the prognosis andradiation-induced lung injury (RILI) vary from patients to patients after TRT. Itindicates that one-size can not fit all in TRT. So, Personalized TRT based ondifferent subtypes of NSCLC has been proposed. Furthermore, biomarker researchhas been considered as the most important way for personalized TRT. This studytested the single nucleotide polymorphisms (SNPs) in lung cancer-related genes andthe levels of autoantibodies against tumor-associated antigens (TAAs)/immunologicregulated molecules (IRMs), and explored any associations between thesebiomarkers and overall survival (OS) and/or radiation pneumonitis (RP) in NSCLCpatients treated with definitive TRT. The aim of this study was to identify usefulbiomarkers for personalized TRT.Methods1. Genetic study (done at the University of Michigan, USA)This study included newly pathologically diagnosed and clinical stage I-IIINSCLC patients treated with definitive TRT. Bioinformatic methods and moleculargenetic techniques were used to test some SNPs in the TGFβ1gene and11DNArepair genes (ATM、ERCC1、ERCC2、ERCC5、MGMT、NBN、TP53、XRCC1、XRCC2、XRCC4、XPC). A total of30SNPs from these12genes were selected asgenetic markers. Matrix-assisted laser desorption/ionization time-of-flight massspectrometer (MALDI-TOF) was utilized to genotypes these SNPs in NSCLCpatients. Kaplan-Meire method, Cox proportional hazards model, receiver operating characteristic (ROC) curve analysis and genetic statistical software package(SPSS19.0and STREE TREE) were used to analyze the association between SNPsand OS/RP.2. Immunologic study (done at Jilin Univerisity)This study included newly pathologically diagnosed and clinical stage I-IIINSCLC patients treated with definitive TRT. A total of7IRMs and TAAs wereselected as antigens in this study, including CD25，FOXP3，TGF-β1，p16，P53，OCT4II and OCT4III. The bioinformatics database and software were used toanalyze their structures and characters. Based on in silico mapping of MHCrestricted epitopes derived from these proteins, linear antigen fragments were thendesigned; the linear peptide antigens were synthezied with a chemical method and anin-house ELISA test was developed to detect circulating autoantibodies in NSCLCpatients. Kaplan-Meire method, Cox proportional hazards model, logistic regressionmodel, ROC curve analysis and genetic statistical software package (SPSS19.0andSTREE TREE) were used to analyze the association between circulatingautoantibodies to TAAs/IRMs and OS/RP.Results1. Genetic analysisOne hundred thirty-four patients who were enrolled in prospective trials withsurvival and lung toxicity data recorded prospectively were evaluated.8patientswere excluded with palliative dose (<60Gy).96patients (76%) underwentconcurrent chemo radiation therapy with the platinum-based regimen. The medianfollow-up time was45months, and the minimum follow-up time was12months.(1) Association between SNPs and OS of NSCLCIn the univariate analysis of clinical characteristics, there were statisticallysignificant associations between OS and sex, tumor volume, KPS or TRT dose. Indetail, patients of female, tumor volume≤118cc, KPS≥80and TRT dose≥70Gywould survive longer. Multivariate Cox analysis in30SNPs showed that5SNPs(ERCC2:rs238406, ERCC1:rs11615, ERCC1:rs3212948, XRCC4:rs9293329,XRCC4:rs2075685) were significantly associated with OS of NSCLC. In detail, patients with genotype of ERCC2:rs238406GT/TT、 ERCC1:rs11615TT、ERCC1:rs3212948CC、XRCC4:rs9293329GA/AA and XRCC4:rs2075685TTwould survive longer. Furthermore, combined analysis of unfavorable genotypes(UFGS) from these5SNPs showed significantly cumulative effects on OS, i.e., thedeath hazard was increased with the numbers of UFGS increasing (0-5). TRT dosesubgroup analysis showed that the effect of SNPs tested on OS was observed only inthe TRT dose <70Gy group (P=0.015) rather than in the TRT dose≥70Gygroup (P=0.112). Interactive analysis by STREE TREE showed that there wasinteractive effect between TRT dose and ERCC2:rs238406or ERCC1:rs11615onOS. ERCC2:rs238406and ERCC1:rs11615could serve as genetic markers for TRTdose <70Gy group and TRT dose≥70Gy group, respectively. ROC analysisshowed that the area under ROC curve (AUC) was less than0.7when clinical factorswere used only. However, the AUC would be more than0.8when clinical factorsand5SNPs were combined together.(2) Association between SNPs and radiation pneumonitisIn the univariate analysis of clinical characteristics, there were statisticallysignificant associations between RP and mean lung dose (MLD) and/or the lungvolume receiving more than20Gy (V20). In detail, patients of MLD≤16.3Gy orV20≤26%would be less likely to develop RP after TRT. Multivariate Cox analysisshowed that2out of30SNPs, XRCC4:rs1478486and XRCC4:rs2075685weresignificantly associated with RP. In detail, patients with genotypes ofXRCC4:rs1478486CC or XRCC4:rs2075685could have higher risk of RP after TRT.Furthermore, combined analysis of unfavorable genotypes (UFGS) from these2SNPs showed a significantly cumulative effect on RP, i.e. an increase in the RPhazard with the numbers of UFGS increasing (0-2). Interactive analysis showed thatthere was an interactive effect between lung dosimetric metrics (MLD/V20) andXRCC4:rs1478486CC or XRCC4:rs2075685on RP. ROC analysis showed that theAUC was0.66when MLD/V20was used only. However, the AUC was more than0.76when MLD/V20and2SNPs were combined together. 2. Immunologic analysisEighty-eight patients who were enrolled in prospective trials with survival andlung toxicity data recorded prospectively were evaluated,10of whom were excludeddue to palliative dose (<60Gy), and61(78%) underwent concurrent chemoradiation therapy with the platinum-based regimen. The median follow-up time was10.3months, and the minimum follow-up time was6months.(1) Association between circulating antibodies to TAAs/IRMs and OS of NSCLCIn the univariate analysis of clinical characteristics, there were statisticallysignificant associations between OS and age or chemotherapy. In detail, patientsaged≤60, who received systemic chemotherapy, could survive longer. MultivariateCox analysis showed that none of circulating autoantibodies to7TAAs/IRMs wassignificantly associated with OS of NSCLC (P>0.05)(2) Association between TAAs/IRMs and radiation pneumonitisIn the univariate analysis of clinical characteristics, there was no clinicalvariable significantly associated with RP. Multivariate logistic analysis of antibodiesto7TAAs/IRMs showed that anti-CD25IgG and anti-OCT4II IgG weresignificantly associated with RP. In detail, patients with a low level of anti-CD25IgG or a high level of anti-OCT4II IgG could have a higher risk of RP after TRT.ROC analysis showed that the AUC was0.59when MLD/V20was used only.However, the AUC would be0.77when MLD/V20and2TAAs/IRMs werecombined together.Conclusion(1) Genetic variants in DNA repair genes are significantly associated with survivalof NSCLC patients treated with definitive TRT. Some SNPs from the DNA repairgenes can serve as biomarkers in predicting survival of NSCLC.(2) Higher dose TRT (≥70Gy) may improve survival in NSCLC patients withunfavorable genotypes of DNA repair genes. However, NSCLC patients carryingfavorable genotypes of DNA repair genes may not benifit from high dose TRT.(3) Genetic variants in DNA repair genes are significantly associated with radiationpneumonitis after TRT. Some SNPs from the DNA repair genes can serve asbiomarkers for prediction of radiation pneumonitis. (4) Circulating autoantibodies against CD25and OCT4II are significantlyassociated with radiation pneumonitis after TRT and can serve as biomarkers forprediction of radiation pneumonitis.(5) A novel genetic and immunological method can be set up based on present workto screen biomarkers for cancer.更多还原