Combination of epithelial-mesenchymal transition and cancer stem cell–like phenotypes has independent prognostic value in gastric cancer

Abstract

Epithelial-mesenchymal transition-related proteins have been suggested to interact with each other in various cancers and be associated with the aggressive behavior of cancer. To demonstrate the clinical significance of epithelial-mesenchymal transition and stem cell-like phenotypes in gastric cancer, we performed immunohistochemistry for 5 epithelial-mesenchymal transition-related proteins, including Snail-1, ZEB-1, E-cadherin, vimentin, and β-catenin, and the gastric cancer stem cell marker CD44 in 276 consecutive primary gastric cancers and 54 matched lymph node metastases. Loss of E-cadherin expression and aberrant expression of vimentin were significantly associated with aggressive clinicopathologic features. The expression of epithelial-mesenchymal transition-related proteins was closely related to each other in gastric cancer. The known gastric cancer stem cell maker, CD44, was significantly associated with the protein expression of Snail-1, ZEB-1, and E-cadherin (P < .05). Univariate survival analysis was performed for the 6 proteins included in this study to find the best combination for predicting patient outcome. Protein expression of Snail-1, vimentin, E-cadherin, and CD44 resulted in the lowest P value using the Kaplan-Meier method (P < .001). This combination of proteins was significantly associated with advanced pT stage, lymph node metastasis, vascular invasion, and undifferentiated histologic type in a high-risk group (P < .001) and predicted disease-free survival independent of pTNM stage and histologic differentiation (P = .029). However, the acquired mesenchymal phenotype of gastric cancer cells at the primary site was restored to an epithelial phenotype in lymph node metastases. A combination of epithelial-mesenchymal transition and stem cell-like phenotypes is an important predictor of aggressive biologic behavior and has an independent prognostic value in predicting outcomes of primary gastric cancer.