【作者】 任景丽；

【导师】 王立东；

【作者基本信息】 郑州大学 ， 内科学， 2011， 博士

【摘要】 1.研究背景贲门腺癌(Gastric cardia adenocarcinoma, GCA)是我国北方特别是河南林州、安阳等地区最常见的恶性肿瘤之一。早期GCA患者5年生存率可高达90%,中晚期患者5年生存率仅10%左右。但临床80%以上就诊的GCA患者均为中晚期。导致GCA预后极差。贲门腺癌的病因不明,遗传因素与环境因素在贲门癌发生过程中的作用仍不清楚。也是导致其死亡率多年无明显改善的主要原因之一。尽管国内外学者多年来对贲门腺癌易感基因进行大量的研究,但因为缺乏高通量检测技术,故仍未取得突破性进展。全基因组关联研究(Genome-wide Association Study, GWAS)技术的问世,正好解决这一技术难题。GWAS是利用高通量基因分型技术分析数以万计的单核苷酸多态性(SNPs)以及这些SNPs与临床表现和可测性状的相关性,对大样本资料进行检测分析,旨在寻找与疾病相关性密切的一组基因。利用GWAS搜寻复杂疾病的易感基因,是学术界公认的目前最有效的方法。自2005年以来,国际上利用GWAS对百余种常见疾病进行研究,包括恶性肿瘤,如前列腺癌、肺癌等以及糖尿病、肥胖、精神疾病进行易感基因的筛选研究,取得一定成绩。在GWAS责门癌方面的研究仅见我们团队自己的报道。本研究通过贲门癌大样本流行病学调查,分析中国汉族人群GCA遗传倾向(家族史)的分布特征和规律。并利用GWAS技术寻找验证贲门癌易感基因,在分子水平探讨中国汉族人群发生GCA的基因多态性改变,并分析吸烟、饮酒、体重指数(BMI)、幽门螺杆菌(Hp)等因素对中国人群GCA发病的影响,为筛选贲门癌变关键候选基因和蛋白、建立高危人群预警和个体化预防的分子指标提供重要的理论依据。2.材料及方法2.1 GCA遗传倾向(家族史)研究2.1.1研究对象共调查贲门癌患者16470例。其中,男性12804例,平均61±9岁；女性3666例,平均61±9岁。所有资料来自河南安阳、林州、新乡等省内30余家医院,河北、广东、四川、云南等省外15余家医院确诊贲门癌的住院病人。2.1.2研究方法采取问卷调查方式,所有结果采用EXCEL软件进行数据录入。采用SPSS17.0统计软件处理,组间比较采用χ2检验(α=0.05)。2.2全基因组关联分析(GWAS)2.2.1研究对象共收集贲门癌及健康对照血标本13779例(男7782例,女5997例),其中病例组2766例(男2144例,女622例,平均年龄61±9岁),对照组11013例(男5638例,女5375例,平均年龄48±14岁)。研究对象来源同本研究第一部分,病例组均经组织学证实,所有患者均未进行过放疗和化疗。对照组来源同上,经胃镜检查排除早期癌。2.2.2研究方法用德国QIAGEN基因组DNA提取试剂盒提取血液样本基因组DNA,准确测定每份待标化样本DNA的浓度,浓度要求15ng-20ng/μl, OD值在1.7-2.0之间。选择Sequenom iPLEX芯片验证。计算样本得率、SNP得率、最小等位基因频率(minor allele frequency, MAF)、哈迪·温伯格(Hardy-Weinberg; HWE)平衡律。对质控后的数据用Plinkl.03软件分析并输出结果,用Cochran-Armitage趋势检验计算P值、优势比(Odds ratio,OR)和95%可信区间(95% confidence interval,95%CI)。吸烟、饮酒、BMI等亚组分析使用SPSS 17.0统计软件处理,进行χ2检验、Logistic回归分析(α=0.05)。2.3免疫组织化学研究2.3.1研究对象取贲门癌组织50例,癌旁不典型增生17例,正常对照组织13例,经常规85%酒精固定、石蜡包埋、切片厚4μm 2张,分别做HE染色和免疫组化染色。一抗为美国Sigma公司提供的兔抗人磷脂酶CE1(PLCE1)多克隆抗体,ABC试剂盒为Vector公司产品。每批实验都设阳性对照和阴性对照。2.3.2统计学方法采用SPSS 17.0统计软件处理,组间的比较采用χ2检验(α=0.05)。3.结果3.1贲门腺癌遗传倾向(家族史)调查结果3.1.1不同性别GCA肿瘤家族史特征贲门癌家族史阳性(FH+)占22%(3669/16470),其中,男性FH+占22%(2870/12804),女性FH+占22%(799/3666),统计学分析性别间无显著差别(P>0.05)。3.1.2高、低发区GCA患者发病年龄与肿瘤家族史特征高发区贲门癌患者发病年龄均数显著小于低发区患者(60.12±9.03 vs.61.12±9.63,P<0.01)；男性GCA患者发病年龄均数比较,高发区亦显著小于低发区(60.16±8.99 vs.61.33±9.63,P<0.01);女性GCA患者中,两组相比无显著差异(59.97±9.16 vs.60.28±10.34,P>0.05)。小于60岁(包括60岁)GCA患者高发区所占比例显著高于60岁以上组(77.8%,6592/8475 vs.74.5%,5956/7995,P<0.01)；男性GCA患者中,小于60岁高发区患者所占比例显著高于60岁以上组(77.4%,5074/6558 vs.73.5%,4589/6246,P<0.01)；女性GCA患者中,两组相比无显著差异(79.2%,1518/1917 vs.78.2%,1367/1749,P>0.01)。高发区患者FH+明显高于低发区(28%,3095/11187 vs.16%,574/3608),统计学分析有显著差别(P<0.01)。3.1.3发病年龄与GCA肿瘤家族史关系分析家族史阳性贲门癌患者发病年龄均数显著小于家族史阴性患者(59.28±8.87vs.60.38±9.28,P<0.01)；男性GCA患者发病年龄均数比较,家族史阳性者亦显著小于家族史阴性者(59.28±8.89 vs.60.52±9.19,P<0.01);女性GCA患者中,两组相比无显著差异(59.32±8.81 vs.59.90±9.60,P>0.05)。小于50岁(包括50岁)GCA患者中FH+者所占比例显著高于50岁以上组(27.0%,599/2215 vs.24.4%,3070/12580,P<0.01)；男性GCA患者中,小于50岁(包括50岁)FH+者所占比例显著高于50岁以上组(27.8%,471/1693 vs.24.4%,2399/9843,P<0.01)；女性GCA患者中,两组相比无显著差异(24.5%,128/522 vs.24.5%,671/2737,P>0.01)。3.2全基因组关联分析结果中国汉族人群13779例样本(2766例病例和11013例对照)中,经GWAS高通量重复验证,筛选出2个新的易感位点：rs2274223和rs13042395。经一系列比对分析后,将rs2274223定位于10q23的phospholipase C epsilon 1 (PLCE1)基因,rs13042395定位于20p13的chromosome 20 opening reading frame 54(C20orf54)基因。3.2.1亚组分层分析：基因多态性和吸烟/不吸烟、饮酒/不饮酒关系分析PLCE1基因rs2274223位点基因型频率病例组和对照组之间χ2检验发现：分布差异非常显著。对基因型进行Logistic回归分析后发现：吸烟/不吸烟、饮酒/不饮酒人群,GG、GA基因型与GCA均具有显著相关。经多因素调整后,吸烟人群和不饮酒人群仍存在显著相关：不吸烟人群和饮酒人群GG基因型仍存在显著相关,GA基因型无显著相关。3.2.2亚组分层分析：基因多态性和BMI关系分析PLCE1基因rs2274223位点基因型在病例组和对照组的Ⅰ级/Ⅱ级BMI人群中进行Logistic回归分析后发现：GG、GA基因型与GCA均显著相关。Ⅲ级BMI人群Logistic回归分析后发现：GG、GA基因型与GCA均无显著相关。3.2.3亚组分层分析：基因多态性和Hp感染关系分析PLCE1基因rs2274223位点基因型在病例组和对照组的Hp感染阳性/阴性人群中进行Logistic回归分析后发现：GG、GA基因型与GCA均显著相关；多因素调整后结果显示：Hp感染阴性人群中GG、GA基因型与GCA的患病风险显著相关；Hp感染阳性人群中GG基因型与GCA发病风险相关,而GA基因型无显著统计学差异。3.2.4 PLCE1基因rs2274223位点基因型与家族史、性别、高低发区、不同年龄人群中进行Logistic回归分析后发现:GG、GA基因型与GCA均显著相关。经多因素调整后,亦有一定的相关性。3.2.5 C20orf54基因多态性rs13042395位点基因型频率病例组和对照组之间χ2检验发现,其基因型分布有显著性差异。在吸烟/非吸烟、饮酒/非饮酒等人群中进行Logistic回归分析,结果显示,该位点基因型与GCA无显著相关。3.3免疫组织化学染色结果正常贲门组织已出现明显PLCE1蛋白表达,但不典型增生(DYS)(88%,15/17)和GCA组织(72%,36/50)中PLCE1蛋白表达阳性率显著高于正常组织(23%,3/13)。PLCE1蛋白阳性表达率与临床病理特征如TNM分期、分化程度和淋巴结转移等关系不明显。4.结论4.1通过大样本的GWAS研究,在国际上首次发现与中国汉族人GCA发病风险显著相关的2个重要SNP易感位点rs2274223、rs13042395,并分别定位于10q23的PLCE1基因和20p13的C20orf54基因上。4.2中国汉族人群PLCE1基因多态性位点rs2274223基因型为GG、GA者发生GCA的风险较AA型高,其中GG基因型是风险因素,GA基因型是条件风险因素。吸烟、不饮酒、Ⅰ级/Ⅱ级BMI、无Hp感染等因素均可能增加GA基因型人群患GCA的风险。4.3Ⅲ级BMI可能是GCA的保护因素。4.4中国汉族人群C20orf54基因多态性位点rs13042395与GCA发病风险有关,但与吸烟、饮酒、BMI等因素之间无相互作用。4.5 PLCE1免疫组化结果提示,PLCE1在贲门癌变过程中可能起重要作用。4.6贲门癌具有明显家庭聚集现象,家族史阳性贲门癌患者平均发病年龄明显小于家族史阴性患者,提示遗传因素在贲门癌发生过程中起一定作用。更多还原

【Abstract】 1. BackgroundGastric cardia adenocarcinoma (GCA) is one of the most common malignant diseases in Northern China, especially in Linzhou and Anyang, Henan Province. 5-year survival rate for early GCA patients can reach 90%, while, in intermediate and advanced GCA,5-year survival is only about 10%. Unfortunately, more than 80% of the GCA patients diagnosed at first time are advanced, so that the prognosis is very poor. For many years, the progress on the GCA research has been gloomy because of the unknown etiology and the unclear effects of genetic and environmental factors on GCA. The susceptibility genes of GCA have been concerned and studied by many scholars throughout the world for many years, but owing to the lack of high-throughput detecting technology, a breakthrough progress is still far from us.The technology of Genome-wide Association Study (GWAS) has just been able to solve such technical challenge. GWAS is to use high-throughput genotyping of thousands of single nucleotide polymorphisms (SNPs) in large scale samples to correlate these SNPs with complex diseases and to identify a group of genes closely associated with the disease. At present, GWAS has been acknowledged in the academe as the most effective method to search for the complex disease susceptibility genes. Since 2005, GWAS has been adopted in the world on more than 100 common diseases, including cancers, such as prostate and lung cancers, diabetes, obesity and mental diseases etc., and certain achievements have been made in identifying susceptibility genes, while GWAS research on GCA is only limited to the report of our own research team.The present study was thus designed to characterize the genetic predisposition (family history) of GCA. Through large sample epidemiological survey and GWAS was further adopted validating the susceptibility genes to GCA. Moreover, the genotyping from GWAS were further analyzed with smoking, alcohol drinking, BMI, and Hp infection. The significance of the study is to provide important molecular clues not only for the screening of candidate key genes and proteins related with gastric cardia carcinogenesis, but also for the establishment of molecular biomarkers for high-risk subject screening and individualized prevention.2. Materials and Methods2.1 Analysis of GCA genetic predisposition (family history)2.1.1 Patients16,470 cases with GCA (12,081 males,3,666 females, with a mean age of 61±9 years) were included in the investigation. All the clinical data of the patients was from over 30 hospitals as Anyang, Linzhou and Xinxiang of Henan Province and 15 hospitals of other provinces as Hebei, Guangdong, Sichuan and Yunnan.2.1.2 Clinical data collectionQuestionnaire method was adopted in this study. All the collected data were entered with EXCEL software and were analyzed by SPSS 17.0. The chi-square test was applied for comparison of different groups (a=0.05)2.2 GWAS validation2.2.1 PatientsThe case-control study was employed in this study. A total of 13,779 samples were applied, including 2,766 GCA patients (2144 males and 622 females, with a mean age of 61±9 years) and 11013 controls (5638 males and 5375 females, with a mean age of 48±14 years). All the samples were from the same places as those stated in Part 2.1.1 and all the cases had not received any radiotherapy or chemotherapy before blood sample collection and had been confirmed as GCA by histopathology. All the controls were performed endoscopic examination to exclude early cancers of the esophagus and gastric cardia.2.2.2 DNA extraction, genotyping and statistical analysisGenomic DNA was extracted by standard procedures using Flexi Gene DNA kits (QIAGEN, Germany) and the concentration was also measured accurately. DNA concentration was normalized to15ng-20ng/μl with 1.7-2.0 OD in replication study. Sequenom iPLEX Chip was adopted to testify.Sample rates, SNP rates, minor allele frequency (MAF), Hardy-Weinberg equilibrium rates (HWE) were calculated. Quality controlled data were analyzed and outputted with Plinkl.03 software. P value, odds ratio (OR) and 95% confidence interval (95% CI) were calculated with Cochran-Armitage trend test. The subgroup analyses of smoking, drinking and BMI were analysis by SPSS 17.0 software so as to perform the chi-square test and Logistic regression analysis (a=0.05).2.3 Immunohistochemical analysis2.3.1 PatientsEighty surgically resected tissue samples were enrolled in this study, including 50 cases with GCA,17 cases with dysplasia tissues and 13 cases with normal controls, respectively. All the tissues were fixed by 85% alcohol, embedded in paraffin.3 slices of 4μm thick were collected from each paraffin block for HE staining and immunohistochemistry.The antibody for PLCE1 was polyclonal rabbit anti-human, American Sigma. ABC kits were provided by Vector Company. Each batch of experiment was equipped with positive control and negative.2.3.2 Statistical analysisThe data were analyzed by SPSS 17.0 software. Theχ2 test was applied for comparison of different groups (P<0.05).3. Results3.1 Analysis of GCA genetic predisposition (family history) 3.1.1 GCA FH positive (FH+) and negative (FH-) by genderThe FH+ rate for GCA was 22%(3669/16470), and GCA FH+ between the male and female showed no significant difference (male vs. female:22%,2870/12804 vs.22%,799/3666) (P>0.05).3.1.2 GCA FH+/FH- in high- and low-incidence areasStatistic analysis showed that GCA FH+ rate in high-incidence areas was much higher than that in low-incidence areas (28%,3095/11187 vs.16%,574/3608; P<0.01).3.1.3 The differences in age of onset and GCA family historyThe familial GCA cases showed a significantly earlier onset age than the sporadic cases (59.28±8.87 vs.60.38±9.28, P<0.01), especially in male (59.28±8.89 vs.60.52±9.19, P<0.01). However, the mean onset age in females was similar between the familial and sporadic cases (59.32±8.81 vs.59.90±9.60, P>0.05).The FH+ rate of GCA patients under or equal to 50 years old was statistically higher than those over 50 years old (27.0%,599/2215 vs.24.4%,2965/12161, P<0.01). In male GCA patients under or equal to 50 years old, FH+ rate was also statistically higher than those over 50 years old (27.8%,599/2215 vs.24.4%, 2965/12161, P<0.01). But in female GCA patients, there was no difference between the FH+ rates of the two groups.3.2 GWAS validation analysisIn 13779 samples of Chinese Han population (2,766 GCA cases and 11,013 controls), after GWAS validation analysis, two SNPs show independent association evidence:rs2274223 and rs13042395. After a series of comparison analysis, including sequencing, one gene at each locus was implicated:PLCE1 at 10q23 and C20orf54 at 20p13.3.2.1 Analysis on smoking/non-smoking, drinking/non-drinking subgroups and genetic polymorphismThe genotype frequency of rs2274223 polymorphism within PLCE1 gene showed statistically significant difference between the case and control groups by chi-square test. The result of logistic regression analysis revealed that in the groups of smokers/non-smokers and drinkers/non-drinkers, the association of GCA with the genotype GG and GA remained significant. After the multi-factored adjustment, the smokers and non-drinkers were still associated with GG. In non-smokers and drinkers, the genotype GG still showed significant association, but without statistically significant differences in genotype GA.3.2.2 BMI and genetic polymorphismRs2274223 polymorphism within PLCE1 gene in the cases and controls with GradeⅠ/Ⅱof BMI was analyzed by logistic regression analysis. The results revealed that the association of GCA with the genotype GA and GG remained significant. In cases and controls with overweight (GradeⅢof BMI), the logistic regression analysis showed that there was no significant association between GCA and genotype GG and GA.3.2.3 Hp+/Hp-subgroups and genetic polymorphismThe result of logistic regression analysis for Rs2274223 polymorphism within PLCE1 gene revealed that the association of GCA with the genotype GA and GG remained significant in Hp+/Hp-of cases and controls. After adjusting gender, the genotype GA and GG remained significant with the risk of GCA in Hp-subgroups, and while the genotype GA showed no statistically significant differences for HP-subgroups.3.2.4 Rs2274223 polymorphism within PLCE1 gene in family history, gender, high-and low-incidence areas and age of cases and controls was analyzed respectively by logistic regression analysis. The result of logistic regression analysis revealed that the association of GCA with the genotype GA and GG remained significant. After Multi-factor adjustment, the genotype GA and GG still remained significant. 3.2.5 Rs 13042395 polymorphism within C20orf54 gene in cases and controls was significantly difference by chi-square test. The result of logistic regression analysis in the cases and controls with smoking/non-smoking, drinking/non-drinking or Grade I/II/III of BMI revealed that the genotype distribution had no significant association with those factors.3.3 PLCE1 protein expression in GCAThe positive rates of PLCE1 protein expression by immunohistochemistry in dysplasia and GCA tissues were 88%(15/17), and 72%(36/50) respectively, which was significantly higher than that of normal cardia 23%(3/13). There was not any clear association between PLCE1 protein expression and the Clinical and pathological features of GCA.4. Conclusions 4.1 We performed a large GWAS in Chinese Han population, and identified a consistent association at 2 loci (rs2274223 and rs13042395) with GCA. One gene at each locus is implicated:PLCE1 at 10q23 and C20orf54 at 20p13. The study suggests that the 2 novel susceptibility genes PLCE1 and C20orf54 are important susceptibility gene for GCA in Chinese Han population.4.2 The stratified analysis for subgroups indicates that SNP rs2274223 genotype GA and GG on PLCE1 shows a higher risk for GCA than genotype AA in Chinese Han population. Smoking, non-drinking, BMI (Ⅰ/Ⅱ) and Hp-may increase the risk of GCA for those with genotype GA.4.3 BMI (Ⅲ) may be a protective factor of GCA.4.4 Rs13042395 polymorphism within C20orf54 gene shows a high risk to GCA in Chinese Han population, but had no significant association with smoking, drinking, and BMI.4.5 The immunohistochemical results suggest that the gene variations of PLCE1 may play an important role in gastric cardia carcinogenesis.4.6 The apparent familial aggregation and earlier onset in males with familiar GCA suggests that genetic factors may play an important role in GCA.更多还原