Identification of cancer/testis-antigen genes by massively parallel signature sequencing

Abstract

Massively parallel signature sequencing (MPSS) generates millions of short sequence tags corresponding to transcripts from a single RNA preparation. Most MPSS tags can be unambiguously assigned to genes, thereby generating a comprehensive expression profile of the tissue of origin. From the comparison of MPSS data from 32 normal human tissues, we identified 1,056 genes that are predominantly expressed in the testis. Further evaluation by using MPSS tags from cancer cell lines and EST data from a wide variety of tumors identified 202 of these genes as candidates for encoding cancer/testis (CT) antigens. Of these genes, the expression in normal tissues was assessed by RT-PCR in a subset of 166 intron-containing genes, and those with confirmed testis-predominant expression were further evaluated for their expression in 21 cancer cell lines. Thus, 20 CT or CT-like genes were identified, with several exhibiting expression in five or more of the cancer cell lines examined. One of these genes is a member of a CT gene family that we designated as CT45. The CT45 family comprises six highly similar (>98% cDNA identity) genes that are clustered in tandem within a 125-kb region on Xq26.3. CT45 was found to be frequently expressed in both cancer cell lines and lung cancer specimens. Thus, MPSS analysis has resulted in a significant extension of our knowledge of CT antigens, leading to the discovery of a distinctive X-linked CT-antigen gene family.

Fig. 1.

Schematic summary of the genes analyzed at each step of this study. Gene numbers in each category are shown in parentheses.

Fig. 2.

Expression of five CT-antigen genes in lung and breast cancer. mRNA expression of the CT genes in non-small-cell lung cancer (X) and in breast cancer (O) was examined by real-time RT-PCR. Each symbol represents one case. The dashed line at 100% indicates the testicular level of expression with which tumor expression levels were compared.

Fig. 3.

CT45 gene family and transcript variants. The location of the six members of the CT45 gene family and their relation to the LOC203522 and SAGE genes on the X chromosome are shown. Transcriptional orientations are indicated by arrows. The three transcriptional variants of CT45 are shown schematically, with boxes indicating exons. Untranslated regions (UT) at the 5′ and 3′ ends are shown as shaded boxes. Translational initiation sites (ATG) and termination codons (*) are indicated.

Fig. 4.

Multiple sequence alignment of the conserved region shared among SAGE, DDX26, LOC203522, and CT45. Identical sequences are shown in black, and conservative changes are shown in gray. The amino acid number in each gene that represents the starting point of the conserved segment is indicated. Although sequence similarities were also identified among SAGE, LOC203522, and CT45 at their N-terminal sequences upstream, these segments were not as conserved and were not seen in DDX26.