Identification of cancer/testis genes by database mining and mRNA expression analysis

Abstract

Cancer/testis (CT) antigens are immunogenic proteins expressed predominantly in gametogenic tissue and cancer; they are considered promising target molecules for cancer vaccines. The identification of new CT genes is essential to the development of polyvalent cancer vaccines designed to overcome tumor heterogeneity and antigen loss. In the current study, a search for new CT genes was conducted by mining the Unigene database for gene clusters that contain expressed sequence tags derived solely from both normal testis and tumor‐derived cDNA libraries. This search identified 1,325 different cancer/testis‐associated Unigene clusters. The mRNA expression pattern of 73 cancer/testis‐associated Unigene clusters was assessed by reverse transcriptase polymerase chain reaction. Three gene products, CT15/Hs.177959, CT16/Hs.245431 and CT17/Hs.178062, were detected only in testis and in tumor tissue. CT15 is equivalent to ADAM2/fertilin‐β. CT16, an uncharacterized gene product, has homology (30–50%) to members of the GAGE gene family and is 89% identical to CT16.2/Hs.293317, indicating that CT16 and CT16.2 are members of a new GAGE gene family. The uncharacterized gene product, CT17, has homology (30%) to phospholipase A1. RT‐PCR analysis showed that CT15 is expressed exclusively in renal cancer, whereas CT16 and CT17 are expressed in a range of human cancers. Real‐time RT‐PCR analysis of newly defined CT genes and the prototype CT antigens, MAGE‐3 and NY‐ESO‐1, revealed low levels (less than 3% of the level detected in testis) of CT15, CT16 and NY‐ESO‐1 in a limited range of normal, non‐gametogenic tissues. This study demonstrates the merits of database mining with respect to the identification of tissue‐restricted gene products expressed in cancer.