Evaluation of malignant and benign gastric biopsy specimens by mRNA expression profile and multivariate statistical methods.

Background: mRNA expression array and multivariate statistical analysis of gastric biopsies can yield insight into the molecular biology basis of local alterations, supporting expression-based identification of morphological alterations.

Methods: From 11 patients with erosive gastritis(EG), 5 with adenocarcinoma (GC), 11 with atrophic gastritis (AG) gastric biopsies were collected, total RNA isolated, T7 amplification and expression analysis of 1047 mRNAs was performed using commercial glass arrays (Clontech, USA). After microarray quality control, applicable data were available from 7 EG, 4 GC, and 5 AG. Multivariate statistical and cell functional analysis were performed. Real-time RT-PCR and immunohistochemistry were used for validation.

Results: GC was characterized by overregulated v-raf, v-erb-a, BCL2-associated- athanogene, immediate-early-response-3, Polo-like kinase, CDK-2, cyclin-C, Pin1 genes, and downregulated ADP-ribosyltransferase, sialophorin and DCC. AG cases had increased PDGF-receptor, TGF-beta-receptor-3, and decreased death-associated-protein-3, beta-1-catenin, topoisomerase-1 levels. In EG upregulation of IGF-receptor-1, CD9, transferrin receptor, integrins, and underexpression of keratin-5, caspase-4 was found. Discriminant analysis could reclassify all samples correctly using four parameters.

Conclusions: mRNA expression array analysis of gastric biopsies yields previously known and new data in the evaluation of local gastric alterations.