| Gastric cancer is one of the most common malignant tumors in the world. In2008,21500new cases of gastric cancer were diagnosed, with730,000deaths fromthe disease. The current5-year survival rate is95%, if the person with early detectionin early stage of gastric cancer. Therefore, early diagnosis and treatment represent akey to improving the prognosis of gastric cancer patients. With the development ofscience and technology to carry out a wide range of various groups, such as thesequence of genome, transcriptome, proteome technology, it is possible for fastfinding tumor biomarkers. But how to find the information about tumor in thecomplex mass data has become the bottleneck of research in the field of cancer.Therefore, bioinformatics has become an indispensable method in cancer research.In the past studies, effective diagnostic biomarkers were detected in tissue orblood, because there were many proteins and peptides and other signal molecules. Itwould really represent human microenvironment in the physiological and pathologicalconditions, so the biomarkers of blood were studied depth. Compared with blood,urine composition is relatively simple and the collection of the urine is easyacquisition and non-invasive. Therefore, the search biomarkers in the urine havebecome a hotspot in the field of cancer research. But there was no substantial breakthrough for a long time to search for tumor markers in urine, the main reason isthe low content of urine protein, and it is easily affected by various factors such asdiet, age, renal function and other factors. In present studies, small range proteinwere detected in the existing proteomics technique, resulting in some measure oflower protein were lost.In this study, we detected the spectrum expression gene in the gastric cancer andnormal gastric tissue with Affymetrix Human Exon1.0ST gene chip. On the basis ofthis, using the SVM-RFE method to establish the expression of the encoded protein,which is a classifier to predict differences in urinary protein. Finally, endotheliallipase protein was detected in gastric cancer in urine by Western blot.1. Screening of gastric cancer related gene expression by microarrayWe detected the spectrum expression gene in the gastric cancer and normalgastric tissue with Affymetrix Human Exon1.0ST gene chip, and molecular functionanalysis involved in biological process was carried out by DAVID database.The data showed that there were2540differentially expressed genes in thegastric cancer tissue, including1984genes were significantly up-regulated,556genesdown-regulated significantly. Up to two fold change differential expression gene were715. The503genes were up-regulated and212genes were down-regulated. The up-regulated genes involved in cell cycle and its regulation, cell motion, inflammation,angiogenesis, nucleic acid metabolism, energy metabolism and activation of the enzyme. The down-regulated genes are mainly involved in digestive, ion transport,steady process, carbohydrate metabolism, regulation system, the reaction of organicmatter, channels and receptor activation.2. Study on gastric cancer urine marks of support vector machine based onForecastingIn order to find the remarkable characteristic urinary protein, classificationmodel was established by SVM, to predict whether the difference expression genesencoding protein can be secreted into urine of gastric cancer. Biological function wasanalyzed by DAVID database.The sensitivity and specificity of classification model is78%and92%respectively. They can distinguish the secretion urine protein and no secretion urineprotein. We analysis and forecast the715difference expression genes encodingprotein, in order to determine whether the protein was secreted into urine. Our datashowed that there were201proteins can be secreted into the urine. Through theanalysis of the literature, comparative gene difference amplitude as well as thefunction enrichment analysis, finally we got71potential urinary biomarkers of gastriccancer, greatly reduced the range of tumor markers. In addition, the antibody arraydata showed that Flt-3ligand, TIMP-1and CEACAM1were substantially suppressedin gastric cancer samples, and one protein (Siglec-5) is more abundant in cancersamples. It was a new and effective way for potential biomarkers of gastric cancer in urine.3. Significance of urine endothelial lipase on screening of gastric cancerAmong these proteins, we selected the LIPG code endothelial lipase protein asdetection object. We detected the EL expression of urine sample, blood, cells andtissue in gastric cancer patients by western blot, we also checked the EL levels in theurine samples of other cancer types and benign lesions to check if EL was specific togastric cancer, to explore the sensitivity and specificity of EL in gastric cancerdiagnosis, providing new detection indicators for clinical screening of gastric cancer.The results showed that the expression levels of EL protein decreasedsignificantly in urine of gastric cancer patients. The expression levels of EL innormal human urine was57, and in90gastric cancer patients, there were71cases(78.9%) no detection of EL protein expression, the low expression or expression ofEL protein were19cases (21.1%), it was a significant difference, P <0.0001. Thevalue of AUC was0.967. The sensitivity and specificity of the assay was79%and100%respectively. There were no significant statistical relationships between the ELexpression levels and age, sex, histologic type, stage and grade, clinical pathologicalcharacteristic. There was no significant different in EL expression levels betweencancerous tissues, cell, blood and healthy subjects. The EL content in urine has noobvious difference in other cancer and benign lesions. These results suggested EL can be regarded as a potential biomarker of gastric cancer. |
PDF Full Text Request