Abstract
The mTOR signaling pathway has been linked to various cancers, but the contribution of alterations in this pathway to clinicopathological characteristics have not been established in gastric cancer. To investigate PIK3CA mutations and the expression of proteins in the PI3K/Akt/mTOR signaling pathway in sporadic gastric cancer. We analyzed PIK3CA mutation and microsatellite instability as well as immunohistochemical expressions of p-Akt, PTEN, p-mTOR, p-4EBP1, p-S6, p-p70S6, and eIF4E in 368 FFPE (formalin-fixed paraffin embedded) tissue from patients with sporadic gastric cancer. Associations between expression and clinicopathologic parameters and patient survival were evaluated. We found PIK3CA mutations in 4 of 173 cases (2.3%). In immunohistochemical analyses, we detected positive p-Akt expression in 22.0% of cases (81/368), negative PTEN expression in 21.5% of cases (79/368), positive p-mTOR expression in 68.6% of cases (243/354), positive p-4EBP1 expression in 58.2% of cases (202/347), positive p-S6 expression in 42.7% of cases (148/347), positive p-p70S6 expression in 51.1% of cases (179/350), and positive eIF4E expression in 78.3% of cases (275/351). In a clinicopathologic analysis, intestinal type was significantly associated with positive p-4EBP1 expression (P < 0.001). In a Kaplan-Meier survival analysis, PTEN loss (P = 0.002) and pS6 positivity (P = 0.043) are significantly associated with reduced overall survival (OS). PTEN loss (P = 0.001), pS6 positivity (P = 0.009), and eIF4E positivity (P = 0.003) are significantly associated with reduced disease free survival (DFS) (disease free survival). In Cox regression multivariate analysis, PTEN loss was an independent factor of reduced time. Alterations of mTOR pathway protein expression are associated with reduced survival in gastric cancer. Significance was noted in the association of pS6 positivity and eIF4E positivity e with reduced survival in univariate analysis and the association of PTEN loss and reduced DFS in univariate analysis as well as multivariate analysis for DFS.