【作者】 闰玮；

【导师】 田亚平；

【作者基本信息】 中国人民解放军军医进修学院 ， 临床检验诊断学， 2010， 博士

【摘要】 肝细胞癌(hepatocellular carcinoma, HCC)是全世界最常见的恶性肿瘤之一。目前肝癌筛查主要依靠血清甲胎蛋白检测和B超检查对高危人群进行监测,这两种方法的联合使用使部分肝癌在亚临床阶段得以诊断。但由于我国肝癌患者30%-40%为AFP阴性(AFP<20ng/dL),即AFP的敏感性、特异性不高。因此寻找新的特异肝癌相关生物标志,对肝癌的早发现、早诊断、早治疗具有重要意义。基因表达谱芯片是近年来迅速发展起来的芯片类型之一,已广泛应用于了解多种疾病的发病机制,寻找疾病相关基因,建立基因诊断模型等领域。本研究首先运用Affy技术平台的人类全基因表达谱芯片技术,通过筛选9例原发性肝细胞癌患者和5例年龄、性别等相匹配的健康对照个体外周血中血细胞基因mRNA表达,发现726个显著差异表达的探针组(Pc<0.05,FC≥2.00),其中103个显著上调,623个显著下调。其中CXCR4, CX3CR1,TLR4, PTPRC, PTGS2, IFNGR1, LYN, CD74, CALR等基因参与免疫过程。此外,许多基因参与转录调节过程,如PFDN5, JUND, SUB1, ZBTB24, NFKBIA, FOS等,PTPRC, PTGS2, RPS17, RPL39, RPS6, RPS24等基因参与生物合成过程。为验证基因芯片的结果,我们扩大实验标本量,另收集40例原发性肝癌患者和21例年龄、性别等相匹配的健康对照个体外周血标本。首先进行两组样本肿瘤标志物和生化指标的检测,并进行相应的统计分析。然后选择基因芯片结果中显著差异表达的15个基因,采用多重基因检测分析技术(GeXP技术)进行样本中血细胞mRNA含量的检测,CXCR4, SPAG9, FOS, ANXA1, CALR, IL8, PFN1, GPC3, HGF在两组间有显著统计学差异(P<0.05),与基因芯片实验结果一致。GZMA由于缺失数据较多,予以删除。根据基因芯片实验结果,在已进行芯片结果验证的基础上,建立原发性肝细胞癌外周血基因诊断模型。进而扩大样本的种类和数量,另收集40例原发性肝癌患者,10例肝硬化患者,16例肝炎患者,32例包括肝血管瘤,肝腺癌等良性肝胆疾病患者和21例年龄、性别等相匹配的健康对照个体外周血标本,并进行五组样本肿瘤标志物和生化指标的检测和统计分析。然后采用GeXP技术进行15个基因在五组样本中血细胞mRNA的检测。结果显示：CXCR4, SPAG9, FOS, ANXA1, CALR, IL8, PFN1, GPC3, HGF, HSPAIB, RPS24,在五组间有显著统计学差异(P<0.05),而PFDN5, GOS2, RPL27在五组间无显著统计学差异(P>0.05)。GZMA由于缺失数据较多,予以删除。根据以上指标在五组间的统计结果,采用K近邻算法,5折交叉确认,并进行100次运行,计算出各组平均准确率,建立五个外周血基因诊断模型,其中以PFN1, CALR, SPAG9, IL8, HGF, FOS, GPC3, ANXA1, HSPAIB共九个指标所建立的诊断模型二对五组样本的平均准确率最高,分别是：0.760,0.980,0.853,0.970,0.936。研究结果表明,该诊断模型能较好的区分肝癌、肝硬化、肝炎、其它组和健康对照五组标本。综上所述,本研究初步阐明原发性肝细胞癌患者外周血细胞基因表达谱,进行了基因芯片结果的验证,并建立了能较好区分五组标本的肝癌患者外周血基因诊断模型,为进一步了解肝癌的发生发展机制,有效提高肝癌早期诊断率奠定了基础。更多还原

【Abstract】 Hepatocellular carcinoma (HCC) is one of the leading causes of cancer death worldwide. At present, HCC screening mainly depends on detection of serum alpha-fetoprotein (AFP) and ultrasonic diagnosis to monitor high risk group. However, with its low sensitivity and nonspecific elevation in nonmalignant hepatic diseases, we need to explore additional new diagnostic biomarkers that could assist HCC diagnosis.Gene chip is a high sensitivity and high-flux technology for gene expression research. One of the most widely used microarray is expression profiling chip, providing a systematic approach to uncover comprehensive information about the transcription profile of diseases, which has found many applications including discovery of gene functions, tumor related genes, drug evaluation, pathway dissection, and classification of clinical samples.In order to study the whole gene transcription profiling of peripheral blood in HCC patients, Nine HCC patients and five health control individuals matched at age and gender were screened with Affy GeneChip HG-U133 plus 2.0. The results shew that there were 726 significantly differential expression probe sets (P<0.05, FC<2.00),103 probe sets of them were significantly up-regulated and 623 were significantly down-regulated. Among them, the genes such as CXCR4, CX3CR1, TLR4, PTPRC, PTGS2, IFNGR1, LYN, CD74, CALR took part in immunologic process. Many genes participated in transcription regulation process, for instance, PFDN5, JUND, SUB1, ZBTB24, NFKBIA, FOS. The genes, PTPRC, PTGS2, RPS17, RPL39, RPS6, RPS24 and so on, took part in biosynthesis process. To validate the results of the GeneChip, the enlarged the experiment samples, including peripheral blood samples of 40 HCC patients and 21 health control individuals matched with the patients were further studied. Meantime the tumor markers and biochemical parameters of the two groupshave been measured, the results shew that significant difference lay in AFP, ALB, AST, ALT, GGT, ALP (P<0.05). Then 15 genes which were significantly differential expression in the GeneChip experiment:CXCR4, SPAG9, FOS, ANXA1, CALR, IL8, PFN1, HSPA1B, PFDN5, GOS2, GZMA, RPS24, RPL27, GPC3 and HGF were selected to analyze mRNA level in the samples by GeXP technology (a analytical technique to detect multiple genes mRNA in one reaction system), with geometric mean of the two housekeeping genes B2M and P-actin as control, then the fluorescence value of the indexes eliminated the value of the control and we got the relative amount of the target genes. The significant statistical difference were found in the indexes of CXCR4, SPAG9, FOS, ANXA1, CALR, IL8, PFN1, GPC3 and HGF (P<0.05), which was in accordance with results of GeneChip experiment. GZMA was deleted because of many absence data.We set up gene diagnostic models of HCC patients according to the GeneChip experiment results and the validated results. The peripheral blood samples of 40 HCC patients,10 hepatic cirrhosis patients,16 hepatitis patients,32 other hyperplasia Liver disease patients such as hemangiomas of liver, adenocarcinoma of liver, and 21 health control individuals matched with the patients, have been collected for validation studies. First, the tumor markers and biochemical indicators of the five groups have been detected, the results shew that significant difference lay in AFP, ALB, AST, ALT, GGT, ALP (P<0.05). Then mRNA level of the 15 genes were detected by GeXP technology. The results shew that significant statistical difference lay in the indexes of CXCR4, SPAG9, FOS, ANXA1, CALR, IL8, PFN1, GPC3 and HGF (P<0.05), which was in accordance with results of GeneChip experiment. According to the statistical results, we calculated the average accurate rate of the five groups with K neighbor algorithm and 5 fold crossing confirmation, and then set up 5 gene diagnostic models, the model two was made up by nine indexes of PFN1, CALR, SPAG9, IL8, HGF, FOS, GPC3, ANXA1, HSPA1B and was the best diagnostic one with the highest average accurate rate for the five group samples:0.760,0.9800,0.853,0.97,0.936.In summary, we screened the whole gene expression profiling of peripheral blood cells in HCC and validated the results of the GeneChip by GeXP technology. The diagnostic model which can better distinguish the five group samples have been developed. This study could provide some information and technology for studying the mechanism of carcinogenesis and raising early diagnostic rate efficiently in HCC patients.更多还原