【作者】 黄文涛；

【导师】 施达仁； 周晓燕； 陆洪芬； 李小秋；

【作者基本信息】 复旦大学 ， 肿瘤学， 2010， 博士

【摘要】 目的：弥漫大B细胞淋巴瘤(Diffuse large B cell lymphoma, DLBCL)是非霍奇金淋巴瘤最常见的类型,也是一类在临床表现,病理形态和遗传学特征方面具有非常异质性的B细胞淋巴瘤。本研究通过对60例弥漫大B细胞淋巴瘤全基因组表达谱的研究,进行分子分型并筛选出差异表达的基因。并且在mRNA和蛋白水平验证筛选的差异基因的表达,进一步探讨弥漫大B细胞淋巴瘤的分子预后标志及免疫组化蛋白水平的临床病理分型。方法：(1)基因芯片表达谱研究：收集60例OCT包埋的弥漫大B细胞淋巴瘤冰冻组织,临床资料及随访资料均相对较齐全,选用Affymetrix公司HG U133Plus 2.0全基因组表达芯片,进行弥漫大B细胞全基因组表达谱芯片的检测和分析。(2)运用实时定量PCR (quantitative real-time PCR)方法进行差异基因的mRNA水平的验证及相关预后的研究：入组标本为基因表达谱分析的弥漫大B细胞淋巴瘤60例进行mRNA水平的验证,另外增加收集2004年1月-2009年9月间复旦大学附属肿瘤医院OCT试剂包埋冰冻组织19例(Hans临床免疫组化分型的6例GCB-DLBCL及13例non-GCB-DLBCL)共79例DLBCL进行相关预后的研究。根据基因芯片表达谱的差异基因筛选结果,用实时定量PCR技术(SYBR greenⅠ)的方法,分别检测11种相关基因,5个GCB-DLBCL表达下调基因(CCND2、BCL-2、FOXP1, IL6、STAT3),其中2个与JAK-STAT信号通路有关(IL6、STAT3);2个GCB-DLBCL表达上调基因(BCL-6、TP53);另外还有4个在基因芯片检测无表达量的变化,但文献报道与预后相关的基因(LM02、FN1、SCYA3、BCL7A)。表达相对量的计算采用2-ΔΔCt法,通过对79例DLBCL的相对mRNA表达量的分析,并与临床资料相对照,探讨弥漫大B细胞淋巴瘤的预后因素。(3)组织芯片结合基因芯片对弥漫大B细胞淋巴瘤病理分型的探讨：收集2000年1月-2009年9月间复旦大学附属肿瘤医院石蜡包埋淋巴结组织335例,包括弥漫大B细胞淋巴瘤,非特指(DLBCL-NOS)271例,滤泡性淋巴瘤(3级)23例,Burkitt淋巴瘤(BL)14例,富于T和组织细胞弥漫大B细胞淋巴瘤(T/HR-DLBCL)11例,间变性大细胞淋巴瘤(ALCL) 6例,淋巴组织反应性增生10例。所有病例的石蜡包埋组织,进行HE染色形态观察,由两位高年资病理医生复片,根据镜下形态和免疫组化指标诊断和分型。结果：第一部分：(1)60例行基因表达谱分析的DLBCL I临床特征：60例DLBCL男性34例,女性26例,男女比例约为1.3:1；年龄分布在14岁至86岁,平均年龄56.8岁；59例有临床随访资料,51例有临床分期资料。按Hans分类标准,GCB病例20例,non-GCB病例40例(其中EB病毒感染6例)。临床特征在患者性别、临床分期之间无明显统计学意义(p>0.05), EBV感染多见ABC-DLBCL患者(p<0.05)。DLBCL组织学类型主要为中心母细胞变型,免疫母细胞变型及少数间变样变型。(2)基因芯片分析结果：60例DLBCL在分子基因水平被分成了三型,经过监督聚类及非监督聚类和文献中分子分型的对照,在分子水平均可分为三型,即GCB-DLBCL、ABC-DLBCL和Type3-DLBCL,其中GCB-DLBCL型20例,ABC-DLBCL型26例,Type3-DLBCL型14例。20例Hans分型为GCB-DLBCL标本中,通过基因芯片分型证实有15例为GCB-DLBCL,2例为ABC-DLBCL,3例为Type3-DLBCL;40例Hans分型为Non-GCB-DLBCL标本中,5例为GCB-DLBCL,24例为ABC-DLBCL,11例为Type3-DLBCL。根据基因芯片分析的结果,GCB-DLBCL诊断符合率为75%,而non-GCB-DLBCL诊断符合率为87.5%,Kappa检验分析,检测结果具有相关性(p<0.01)。另外筛选出2,170个差异基因,进行了相应的功能分析,在pathway功能分析(KEGG)中发现JAK-STAT3信号通路的基因在ABC-DLBCL较GCB-DLBCL上调。应用基因表达谱在分子水平分为GCB-DLBCL, ABC-DLBCL和type3-DLBCL三组标本和Hans免疫组化分析后在蛋白水平分为GCB和Non-GCB二组标本,分别统计各组病例的无病生存率(Disease-free survival,DFS),应用Kaplan-Meier方法分别进行生存分析,分子分型GCB-DLBCL ABC-DLBCL和type3三组标本和蛋白分型GCB和Non-GCB二组标本均无明显统计学差异(p>0.05)。第二部分：(1)79例DLBCL行实时定量PCR分析的DLBCL临床特征：男性46例,女性33例,男女比例约为1.4:1；年龄分布在14岁至86岁,中位年龄58岁；78例有临床随访资料,69例有临床分期资料。按Hans免疫组化分类标准,GCB-DLBCL病例26例,non-GCB-DLBCL病例53例,患者的性别及临床分期无明显统计学差异(p>0.05)。(2)分别将11个挑选的基因进行临床预后研究,发现BCL-2、IL6、STAT3、BCL-6、TP53、LMO2、FN1、SCYA3、BCL7A共9个基因和预后无明显相关(p＞0.05)；而CCND2, FOXP1和临床分期有提示预后的意义(p＜0.01),但Cox回归模型分析后发现仅有临床分期对DLBCL生存有显著影响(p<0.05)。临床特征方面,患者的年龄和性别经生存分析与预后无关(p>0.05),而临床分期与肿瘤的预后明显相关(p<0.01)第三部分：(1)335例组织芯片淋巴瘤患者临床特征：271例弥漫大B细胞淋巴瘤,非特指(DLBCL-NOS)男性151例,女性120例,男女比例约为1.3:1；年龄分布在9岁至87岁,中位年龄58岁；按Hans免疫组化分类标准,GCB病例65例,non-GCB病例206例。(2)依据免疫组化标记在DLBCL的表达情况,通过不同组间的统计学分析,发现免疫组化分型(IHC分型)中在组间表达有差异的基因有CD10、BCL-6、MUM1和Cyclind2 (p<0.05),基因表达谱分型(GEP分型)中在组间表达有差异的基因有CD10、BCL-6、Cyclind2 (CCND2)和CD44(p＜0.05)。在Hans免疫组化分型的基础上,增加Cyclind2和CD44抗体标记,可提高DLBCL临床病理分型和诊断的准确性。结论：(1)应用Affymetrix全基因芯片表达谱分析了60例弥漫大B细胞淋巴瘤。分别通过非监督聚类、监督聚类及文献筛选基因的聚类分析,均能在分子水平将DLBCL分成明确的三型：ABC-DLBCL、GCB-DLBCL和type3-DLBCL。(2)经基因和蛋白水平的验证,基因表达谱能较准确用于DLBCL的分子分型及差异基因的筛选(3)通过对11种与预后相关基因的生存分析,CCND2, FOXP1在本研究中具有一定的预后提示作用,但可能不能成为独立的预后指标。临床分期对DLBCL生存有显著影响(p<0.05)(4)在Hans免疫组化分型的基础上,增加Cyclind2和CD44抗体标记,可提高DLBCL临床病理分型和诊断的准确性。即Cyclind2和CD44阳性更倾向于GCB-DLBCL; Cyclind2和CD44阴性则要考虑有无Non-GCB-DLBCL的可能更多还原

【Abstract】 Objective:Diffuse large B cell lymphoma (Diffuse large B cell lymphoma, DLBCL) is the most common non-Hodgkin’s lymphoma worldwide. It is clinical, morphologically and genetically a heterogeneous group of tumors composed of large B-cell. We applied the method of microarray to tumor classification and identify the different expression of genes in 60 patients with diffuse large B cell lymphoma. In additional, we will verify significent genes within this set of microarray identified candidate genes in the level of mRNA and protein in 60 cases of diffuse large B cell lymphoma.Methods:(1) genome-wide gene expression analysis of Diffuse large B cell lymphoma:60 patients of DLBCL with OCT-embedded frozen tissue cell lymphoma, clinical data and follow-up data are relatively complete, Affymetrix HG U133 Plus 2.0 whole genome expression chip was selected, for diffuse large B cell whole genome microarray detection. (2) different genes and related outcomes of validation with the method of quantitative real-time PCR in the levels of mRNA:60 cases of diffuse large B cell lymphoma which have analysised by of gene expression profiles, and the other 19 cases of DLBCL OCT embedded frozen tissue from Fudan University Cancer Hospital tissue bank between January 2004 to September 2009, including six cases of GCB-DLBCL and 14 cases of non-GCB-DLBCL with the method of immunohistochemical analysis and the method of quantitative real-time PCR(SYBR green I),11 genes selected by expression of gene expression profile (GEP)were detected, of which five genes was down-regulated expression in GCB-DLBCL (CCND2, BCL2, FOXP1, IL6, STAT3), including 2 genes of JAK-STAT signaling pathway (IL6, STAT3); 2 genes was up-regulated expression in GCB-DLBCL (BCL6, TP53); 4 genes without any change with the result of gene chip analysis, but they have prognosis significant in some papers (LM02, FN1, SCYA3, BCL7A),and using the method of 2 -ΔΔCt to calculated the relative expression. We will explore a new classfication of diffuse large B cell lymphomaand followed by multiple analysis of gene chips to selectable markers and a number of possible genes associated with prognosis in contrast with the clinical data. (3) Tissue microarray associated with gene chip to analysis the pathological classification of diffuse of diffuse large B cell lymphoma: 335 cases of paraffin-embedded lymph node tissue collected in January 2000 to September 2009 in the Fudan University Cancer Hospital, including 271 cases of diffuse large B cell lymphoma, unspecified (DLBCL-NOS),23 cases of follicular lymphoma (grade 3),14 csases of Burkitt lymphoma (BL), 11 cases of T-cells/histiocyte rich diffuse large B cell lymphoma (T/HR-DLBCL),6 csases of anaplastic large cell lymphoma (ALCL),10 cases of reactive lymphoid hyperplasia in. All cases with paraffin embedded and HE staining, their patterns were observed by two pathologists according to the analysis by microscopic morphology and immunohi stochemi stry.Results:Part 1:(1) clinical characteristics of 60 patients of DLBCL by gene expression profiling analysis:60 patients with DLBCL including 34 males and 26 females, male to female ratio was about 1.3:1; age from 14 years to 86 years, mean age was 56.8 years; 59 patients have the clinical stage data and 51 patients have clinical follow-up data with the Hans classification, including 20 cases of GCB,40 cases of non-GCB (including 6 cases with EB virus infection). There are no significant between the group of different clinical features, gender and clinical stage (p>0.05), EBV infection is more common in ABC-DLBCL patients (p<0.05),6 cases of DLBCL with infection with EBV with there is not any the history of clinical immunedeficiency, and the majority (five cases) of EBV infection cases the age more than 50 years, so pathological diagnosis for the elderly diffuse large B cell EBV-positive lymphoma. the major histological type is the center cell variant, immunoblastic variant, and a few anaplastic variations. (2) microarray analysis:using the method of gene expression profile,60 cases of DLBCL has been divided into three types, through the supervised clustering and unsupervised clustering and 38 genes from literature to supervised clustering for control, so 60 cases of DLBCL can be divided into:GCB-DLBCL, ABC-DLBCL and Type3-DLBCL, including 20 cases of GCB-DLBC1 type,26 cases of ABC-DLBCL-type,14 cases of Type3-DLBCL there are total of 2,170 different genes through analysis of gene expression profile. In addition, with the analysis of GO and the method pathway (KEGG)found that JAK-STAT3 signaling pathway genes were much significant in the ABC-DLBCL than in GCB-DLBCL type. moreover, gene chip expression analysis by clustering methods in level of molecule can divided the other two types:EBV+DLBCL and EBV-DLBCL separately.Part 2:(1) 79 case of DLBCL with the method of quantitative real-time PCR analysis:The DLBCL group included 46 male and 33 female, male to female ratio of about 1.4:1 in 79 cases of DLBCL, age from 14 years to 86 years, median age of the patients with DLBCL was 58 years; 78 cases of DLBCL have clinical follow-up data,69 case of DLBCL have clinical stage data. By Hans immunohistochemical classification, there are 26 cases GCB-DLBCL types,,53 cases of non-GCB-DLBCL. There was no significant between the type of gender and clinical features (p>0.05). (2) 11 genes selected from gene expression profile,we found that CCND2, FOXP1, and clinical stage has a significant difference in prognosis (p<0.05).Part 3:(1) 335 cases of lymphoma in patients with the method tissue microarray:diffuse large B cell lymphoma, unspecified (DLBCL-NOS) 271 cases,151 cases was male, female 120 cases, male to female ratio of about 1.3:1; Age distributed in 9 to 87 years, median age was 58 years; by Hans immunohistochemical classification,65 case of GCB-DLBCL,206 cases of non-GCB-DLBCL (2) Based on immunohistochemical expression of the DLBCL, there are statistical significantin of the expression of CD10、BCL-6、MUM1 and Cyclind2 with the methods of IHC (p<0.05), In addition, there are also statistical significantin of the expression of CD10, BCL-6, Cyclind2 and CD44 (p<0.05) with the methods of GEP (p<0.05). It can Improve the accuracy of pathological diagnosis when add the addition antibody of Cyclind2 and CD44.Conclusion:(1) total 60 cases of diffuse large B cell lymphomapplication with analysis by Affymetrix gene chip expression: using unsupervised clustering, supervised clustering and cluster analysis with the significant gene selected from the literature separately, DLBCL can be divided into 3 different types in the level of molecular, that is ABC-DLBCL, GCB-DLBCL and type3. (2) gene expression profile are more accurate than the level of mRNA and protein in DLBCL classification. (3) CCND2, FOXP1 have a little significant in prognosis, but they are not the independent prognostic indicators. (4) Hans immunohistochemistry in the classification based on the increase the antibody of Cyclind2 and CD44 can improve the diagnosis accuracy of DLBCL. it more likely to belong to GCB-DLBCL when Cyclind2 and CD44 protein are positive; whereas Cyclind2 and CD44 protein are negative, it more likely to consider the possible Non-GCB-DLBCL diagnosis.更多还原