Identification Of Gastric Cancer-related Genes By Multiple High Throughput Expression Profile Analysis

ObjectiveTo identify gastric cancer-related genes by combined multiple high throughput expression profiler data mining and to explore the feasibility and method of incorporating the Internet-available bioinformatic databases to discover gastric cancer-related genes. Determining gastric cancer-related genes that are commonly deregulated in different tumor types may facilitate identification of targets for gastric cancer diagnoses and therapeutic treatments.Methods1. A pilot study was performed using both EST and SAGE vNorthern to analyze the expression of 10 known genes related to gastric cancer. Compare vNorthern analysis result with known expression pattern of the genes to validate the method for finding gastric cancer-related genes.2. Using a data-mining tool named Digital Differential Display (DDD) from the UniGene database at the NCBI web site, genes with differential expression levels between gastric cancer and their counterpart normal tissues were obtained for further analysis.3. cDNA digital gene expression displayer (DGED) based EST data from the Cancer Genome Anatomy Project (CGAP) were used to analyze differential expression gene between benign and malignant gastric tissues. Genes with >2 folds differential levels were obtained.4. SAGE digital gene expression displayer (DGED) based SAGE data from the Cancer Genome Anatomy Project (CGAP) were used to analyze differential gene expression between benign and malignant gastric tissues. Genes with >two-fold difference were obtained.5. The genes from DDD, cDNA DGED and SAGE DGED were further analyzed by vNorthern and differentially expressed genes were obtained only accordingto p<0.05 both in EST and SAGE vNorthern.6. Compare the genes from vNorthern with differentially expressed genes produced by gastric microarray data from Stanford Microarray Database and co-differential expression genes were retrieved.7. Five genes of findings including ANXA1, AGR2, ANXA10, PSCA and MSMB were analyzed by RT-PCR in 38 paired human gastric cancer tissue samples. ANXA1 were further analyzed by immunohistochemistry in 38 normal > 22 precancerous and 38 cancer tissue samples.Results1. Of 10 known gastric cancer genes, expression pattern of 8 genes are consistent between vNorthern analysis result and known actual state.2. Digital Differential Display (DDD) retrieved 165 differential expression between benign and malignant gastric tissues. Of them 27genes is up-regulated and 128 genes is down-regulated in gastric cancer tissue.3. cDNA digital gene expression displayer (DGED) retrieved 286 differential expression between benign and malignant gastric tissues. Of them 93 genes is up-regulated and 193 genes is down-regulated in gastric cancer tissue.4. SAGE digital gene expression displayer (DGED) retrieved 201 differential expression between benign and malignant gastric tissues. Of them 73 genes is up-regulated and 108 genes is down-regulated in gastric cancer tissue.5. All genes from DDD, cDNA DGED and SAGE DGED were further analyzed by vNorthern, and 45 genes were obtained according to p<0.05 both in EST and SAGE vNorthern.6. Of the 45 genes ,12 genes were found differentially expressed in microarray data from Stanford Microarray Database.7. ANXA1 is up-regulated, but AGR2, ANXA10, PSCA and MSMB are down-regulated confirmed by RT-PCR in 38 paired human gastric cancer tissue samples. Immunohistochemistry showed the positive immunostaining rate for ANXA1 protein had an increasing tendency from normal group (10.53%), precanerous (40.91%) and cancer(73.68%) with P<0.05. The expression of ANXA1 protein and mRNA was significantly higher in diffuse-type gastric cancer than in intestinal-type gastric cancer (93.75% vs. 54.55%, P<0.05).ConclusionsCombined multiple high throughput expression profiler data mining is an effective strategy of data mining for gastric cancer-related gene identification. Byshowing a gradually increase of ANXAl expression in normal, precancerous and malimant stomach cancerous gastric tissues our findings suggest that the increase of annexin I expression occurs early in gastric tumorigenesis. ANXAl overexpression may serve as a useful potential marker of gastric cancer.