Prediction and analysis of HLA-A2/A24-restricted cytotoxic T-lymphocyte epitopes of the tumor antigen MAGE-n using the artificial neural networks method on NetCTL1.2 Server

Cancer immunotherapy has become one of the most important therapeutic approaches to cancer in the past two decades. Tumor antigen-derived peptides have been widely used to elicit tumor-specific cytotoxic T lymphocytes (CTLs). Antigen-specific CTLs induced by MAGE-derived peptides have proven to be highly efficacious in the prevention and treatment of various types of tumor. MAGE-n is a new member of the MAGE gene family and has been shown to be closely associated with hepatocellular carcinoma. It is highly homologous to the MAGE-A gene subfamily, particularly to MAGE-3 (93%). MAGE-n-derived peptide QLVFGIEVV is a novel HLA-A2.1-restricted CTL epitope that induces MAGE-n-specific CTLs in vitro. Identification of these CTL epitopes may lead to clinical applications of these peptides as cancer vaccines for patients with MAGE-n(+)/HLA-A2(+) tumors. In the present study, HLA-A/A24-restricted CTL epitopes of antigen MAGE-n were predicted using the NetCTL1.2 Server on the web, COMB >0.85. The results showed that the NetCTL1.2 Server prediction method improved prediction efficacy and accuracy. Additionally, 8 HLA-A2- and 9 HLA-A24-restricted CTL epitope candidates (nonamers) derived from the tumor antigen MAGE-n were predicted. These nonamers, following identification via experimentation, may contribute to the development of potential antigen peptide tumor vaccines.