A census of predicted mutational epitopes suitable for immunologic cancer control.

The adaptive immune system can protect against spontaneously arising tumors, and the potential exists to reduce cancer incidence by priming adaptive immune responses with vaccines. Immunologic cancer control has been implemented for cancers caused by infectious agents, but not for spontaneous cancers caused by mutation. This is largely due to the high cost of preventative clinical trials and the lack of validated tumor epitopes. Here we evaluate, computationally, all known somatic mutations in human tumors for their antigenic potential. All possible human leukocyte antigen (HLA) class I presented peptides containing recurrent somatic cancer mutations with frequency > 5% were screened by three independent epitope prediction algorithms (SYFPEITHI, BIMAS, and IEDB). Using stringent filters, a total of 20 genes, 35 mutations, and 159 candidate epitopes were identified, each presented by up to four distinct HLA class I alleles. The top-ranking gene from our survey was KRAS, which figures prominently because there are frequent hotspot mutations in numerous, prevalent cancers, and mutant peptides are predicted to be presented by several common HLA alleles. From our data, we estimate that prophylactic vaccination could provide meaningful levels of prevention of tumors associated with common recurrent mutations. (c) 2010 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.